THE ECOLOGY OF ANAPLASMA PHAGOCYTOPHILUM AND THE BLACKLEGGED TICK, IXODES SCAPULARIS IN THE UPPER MIDWEST, U.S.A. By Vishvapali Chandrika Kobbekaduwa A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Comparative Medicine and Integrative Biology – Doctor of Philosophy 2023 ABSTRACT Ixodes scapularis, commonly known as the blacklegged tick (or deer tick), is a medically important tick species that is spreading multiple diseases in the eastern USA. It is especially important as the vector for transmitting the Lyme disease pathogen, Borrelia burgdorferi and anaplasmosis pathogen, Anaplasma phagocytophilum (Ap). With factors such as habitat restoration, increasing deer densities and climate change, the geographic range of I. scapularis has been expanding. Most of the disease cases are reported from two major foci: the Upper Midwest and the Northeast. With the expansion of I. scapularis it is important to study the ecology of I. scapularis and its associated pathogens. The overall goal in this dissertation was to study the ecology of a tickborne disease and its vector. There are two research chapters and to help understand the context and importance of the research topics, there are two literature review chapters, one to review Ap and one to review Species distribution models (SDMs) with particular attention paid to literature on I. scapularis SDMs. The first research chapter examines the host ecology of A. phagocytophilum at a highly endemic site for I. scapularis in the Upper Midwest. I estimated that eastern chipmunks have relatively greater realized reservoir competence than the white-footed mouse but considering the overall contribution to the enzootic cycle of A. phagocytophilum, white-footed mice may play a larger role because they feed a higher proportion of larvae. Most questing nymphs and all the hosts captured that were infected with A. phagocytophilum were infected with the human pathogenic strain of A. phagocytophilum, Ap-ha. This means that if humans and/or companion canines are bitten by an A. phagocytophilum-infected tick at this field site, there is a high risk that of disease. I found the phenology patterns of infection prevalence of hosts, on-host larvae, and the density of infected nymphs follow that of the phenology patterns of questing nymphs and larvae. Blood and biopsy samples can be used for assaying A. phagocytophilum, but I suggest conducting xenodiagnoses experiments to determine empirically the length of transmission of A. phagocytophilum by each tissue type into larvae. Conclude with what is novel and/or important about the findings. The second research chapter centers on developing species distribution models for I. scapularis in Michigan. In this chapter I looked at the environmental predictors that were important in determining where I. scapularis would occur within Michigan and where suitable habitats for the occurrence of I. scapularis in Michigan are found. I developed models for the Upper and Lower Peninsula. I used two different modeling methods, a logistic regression and a machine learning technique based maximum entropy modeling. For both peninsulas environmental predictors related to temperature, humidity, presence of maple, beech, birch forest types, presence of white- tailed deer, soil moisture, and soil clay content. In the Lower Peninsula, most of the southern regions were considered to have suitable habitat for the occurrence of I. scapularis, while the northern region in the Lower Peninsula had the least suitable habitats for I. scapularis occurrence. In the Upper Peninsula, central southern regions as well as regions along the Lake Superior had suitable habitats for I. scapularis occurrence, while there were pockets of least suitable habitats across the peninsula. Future studies should develop a species distribution model based on the current distribution of I. scapularis in Wisconsin and project it onto Michigan to extrapolate where suitable habitats are found in Michigan. A comparison of that model with the ones we have developed can help us to understand better the invasion process of I. scapularis in Michigan. Furthermore, continuing surveillance efforts in the currently predicted least suitable regions will determine if those habitats really are not suitable for I. scapularis or if it just has not gotten there. Conclude with what is novel and/or important about the findings. This dissertation is dedicated to my family and friends. Thank you for believing in me and supporting me. iv ACKNOWLEDGEMENTS I am grateful for the Midwest Center of Excellence for Vector Borne Diseases, the Hal and Jean Glasson Foundation, The Michigan State University Graduate School Research Fellowships, and especially the Comparative Medicine and Integrative Biology program for providing the financial assistance for my training. For research funding, I am grateful to the Michigan Lyme Disease Foundation (Ms. Linda Lobes), the National Science Foundation, the Centers for Disease Control and Prevention, the Michigan Department of Health and Human Services, and The Michigan State University College of Veterinary Medicine Endowed Research Funds. I also would like to acknowledge my appreciation to the Michigan Department of Natural Resources, Kellogg Biological Station, Fenner Nature Center, Ft. McCoy Army Installation, and many other land managers for allowing us to collect specimens. First and foremost, I would like to thank my advisor Dr. Jean Tsao for being not just my mentor but also a friend over the years. I appreciate her dedication in providing a sound mentorship, in helping me develop as a scientist and a person I am today. She has always supported me and encouraged me to think outside the box and saw my potential such that it helped to push myself to limits I did not know was achievable. I thank her for having faith in me especially during COVID-19 pandemic year in 2020, when I wanted to change the direction of my research. I am lucky to have gotten the opportunity to learn not just about science, but about life in general from her. I am especially grateful for the time she was there for me when I fell and broke my arm and had to undergo surgery three months before I was going to defend. She supported me during that hard time in my life and was like my family away from home. Thank you, Jean, for not just being my mentor but my colleague and my friend. I appreciate your patience, kindness, and the care you extend to all your students and making yourself always accessible and helping them through their v struggles and fears. I would also like to thank my guidance committee members for their support encouragement and advice over the years, Dr. Edward Walker, Dr. Jennifer Owen, and Dr. Shannon Manning. To Dr. Edward (Ned) Walker, I appreciate all the wealth of knowledge you have always shared during our lab meetings and during our conversations over the years. I really enjoyed those conversations where each time I learnt something new, be it about mosquitoes, ticks, or diseases. To Dr. Jennifer (Jen) Owen, I really enjoyed taking disease ecology with you and learning about mist netting and birds; I appreciate all the advice and support you have given over the years. To Dr. Shannon Manning, I appreciate all the advice and support you have given me. I enjoyed taking the pathogenesis class, which made me think about my research study in a different perspective. I appreciate the dedication and time you spend on your students and the fact that you push students to think outside the box. I would also like to thank Dr. William Miller who jumped in to help me with my species distribution modeling chapter. Thank you, Will, for all your words of encouragement, advice and all the feedback when I first started to get into spatial modeling. I appreciate the foundation you have given me to pursue this interest in my future endeavors. I would also like to thank the Comparative Medicine and Integrative Biology program and especially Dr. Vilma Yuzbasiyan-Gurkan for welcoming me into this amazing program and helping me to discover what research area I really enjoy doing, when I came into this program without a any idea. She helped me orient myself to really know what I like and enjoy doing. Vilma thank you so much for giving me the foundation to the scientist I have become today. I would also like to thank Dr Colleen Hegg and Dimity Palazzola for their support and always being available to me whenever I had questions regarding the Comparative Medicine and Integrative Biology program. vi Over the years I have really enjoyed working in Jeans and Ned’s lab. I appreciate all the wonderful lab mates who welcomed me into the lab and helped me with setting up the experiments and all the fun times we had. I thank Seungeun Han who really helped me get set up in the lab and being so welcoming when I first started. You were always there whenever I had questions when I first started out, and you have instilled in me good lab practice which I will carry on into the future. I also would also like to thank my lab mates Megan Porter, Kim Fake, Genevieve Pang, Michelle Volk, Peter Fowler, and Joseph Pastori in Tsao lab for all the advice, support and the friendship given over the years. I would also like to thank Amanda Dolinski in the Owen Lab; and Jennifer Kirk, John Bosco Keven and Rex Mbewe in the Walker lab for their feedback and support in the lab. Finally, I would like to thank all the field assistants who helped with the Michigan surveillance tick collections over the years: Shelly Gleason, Maya Regalado, Karana Wickens, Jake Kryda, Michael Orbain, Adam Talbot, Allison Yackley, Haile Waters, Autumn Dunivant, Dan Houvener, Allison Luchenbill, Sarah Nguyentran, Belinda Wilson, and Lauren Quatroche. I also want to thank all the students who collected all the specimens and data in all prior years, especially at Fort McCoy, Wisconsin (Isis Arsnoe, Lydia Kramer, Anthony Scholze, Seungeun Han) and other sites in the Lyme Gradient project. Finally, none of this would be possible if not for my family and friends. I would like to thank my parents Tikiri Kobbekaduwa and Sujampathika Kobbekaduwa for always supporting and standing by my decisions and believing in me. I thank my aunt and uncle Kamani and Nihal Wijeysundara in Toronto, who first supported me and helped me settle in Lansing when I moved here in 2015. I would especially like to thank Gayan Pathirana for being my best friend and my partner. Gayan (Natta) thank you so much for all the times you have been standing by my side supporting me and pushing me forward whenever I was ready to give up. I know that I was vii impossible at times but thank you for believing in me and helping me realize my potential. I love you. My PhD. journey would not have been possible if not for you, Natta, standing behind the scenes, cheering me on, always pushing me to try new things, telling me to face my fears head on, and most of all for telling me never to give up and that at the end of the day, everything will work out. Another huge thank you goes out to Thilani Anthony. You were not just my roommate, but you were my sister from another mother, my best friend, my cheering squad, and my foodie partner. Deciding to become roommates with you was one of the best decisions I made after moving to the US. I appreciate all the times you cheered me on and all the amazing advice, encouragement and support you have given me over the years, to help me believe in myself and push harder to achieve my PhD, even though all the adversity I had faced. You have inspired me through the years and continue to do so with your courage and perseverance. Finally, I would like to thank all my friends in East Lansing - especially thank you for being there for me when I fell and broke my arm. You guys were not just friends but my family away from family, taking care of me and helping me through my surgery and making me smile even when I was at my lowest point. I made it finally after all those years struggling through so thank you viii TABLE OF CONTENTS CHAPTER 1: INTRODUCTION ................................................................................................... 1 BIBLIOGRAPHY ....................................................................................................................... 7 CHAPTER 2: ANAPLASMA PHAGOCYTOPHILUM: A REVIEW............................................ 11 ABSTRACT .............................................................................................................................. 11 INTRODUCTION ..................................................................................................................... 12 BIBLIOGRAPHY ..................................................................................................................... 29 CHAPTER 3: THE ECOLOGY OF ANAPLASMA PHAGOCYTOPHILUM AT A SITE IN THE UPPER MIDWEST USA (FORT MCCOY, WISCONSIN)........................................................ 44 ABSTRACT .............................................................................................................................. 44 INTRODUCTION ..................................................................................................................... 47 MATERIALS AND METHODS .............................................................................................. 48 RESULTS .................................................................................................................................. 54 DISCUSSION ........................................................................................................................... 74 BIBLIOGRAPHY ..................................................................................................................... 90 APPENDIX ............................................................................................................................. 100 CHAPTER 4: SPECIES DISTRIBUTION MODELING AND APPLICATION TO IXODES SCAPULARIS .............................................................................................................................. 107 ABSTRACT ............................................................................................................................ 107 INTRODUCTION ................................................................................................................... 109 BIBLIOGRAPHY ................................................................................................................... 125 CHAPTER 5: SPECIES DISTRIBUTION MODELING OF THE BLACKLEGGED TICKS (IXODES SCAPULARIS) IN MICHIGAN ................................................................................. 133 ABSTRACT ............................................................................................................................ 133 INTRODUCTION ................................................................................................................... 135 MATERIALS AND METHODS ............................................................................................ 142 RESULTS ................................................................................................................................ 152 DISCUSSION ......................................................................................................................... 170 BIBLIOGRAPHY ................................................................................................................... 184 APPENDIX ............................................................................................................................. 197 CHAPTER 6: CONCLUSION ................................................................................................... 201 BIBLIOGRAPHY ................................................................................................................... 209 ix CHAPTER 1: INTRODUCTION In modern times there is an increasing trend in vector-borne diseases. Globally vector- borne diseases account for 17% of all the world’s infectious diseases claiming around 700,000 lives annually (World Health Organization, 2017). In the United States alone, nearly 650,000 cases of vector-borne diseases were reported to the Centers for Disease Control and Prevention (CDC) from 2004 – 2016, out of which 75% of the cases were attributed to tickborne diseases (Rosenberg et al., 2018). In North America, medically and veterinary important ticks have become an immense disease burden on humans and livestock (Berrada and Telford, 2009; Pérez de Leon et al., 2012; Miller, Farnsworth and Malmberg, 2013; Rosenberg et al., 2018; Wisely and Glass, 2019; Rodino, Theel and Pritt, 2020; Eisen and Paddock, 2021). The increase in ticks and tick-borne diseases in North America has occurred largely due to the recent geographic expansion of the range of ticks (Sonenshine, 2018a; Tsao et al., 2021) Lyme disease, a tick-borne disease, is the most common vector-borne disease in the United States reported annually. Nearly 82% of tickborne diseases are attributed to Lyme disease (Rosenberg et al., 2018). Between the periods of 2010 – 2018, nearly 470,000 cases of Lyme disease were diagnosed in the US with annual number of cases reported ranging between 30,000 – 40,000. (Kugeler et al., 2021) It has been proposed that the annual disease burden of Lyme disease could be estimated to be as much as $1 billion dollars with an average cost of $1200 per patient (Hook et al., 2022). In the US there are two major foci of Lyme disease incidence - the upper Midwest, and the Northeast. These foci lie within the distribution of the vector of Lyme disease in the eastern USA, the blacklegged tick (Ixodes scapularis), which has been expanding for more than a half century (Eisen and Eisen, 2018a; Gardner et al., 2020; Burtis et al., 2022). Ixodes scapularis is also responsible for transmission of several other bacterial pathogens 1 including Anaplasma phagocytophilum, which causes granulocytic anaplasmosis, Borrelia miyamotoi, which causes relapsing fever spirochete, and Ehrlichia muris eauclarensis, which causes ehrlichiosis. Ixodes scapularis is also responsible for transmitting a parasitic organism, Babesia microti, which causes babesiosis, as well as a virus, Powassan virus, which causes an encephalitis. These pathogens and associated diseases are also found mainly in the upper Midwest and Northeast (Fleshman et al., 2021). With the spread of the I. scapularis, the diseases associated with this tick are also on the rise in the US (Eisen and Eisen, 2018a; Rosenberg et al., 2018) Distribution of I. scapularis in Michigan Ixodes scapularis have been expanding in range in both the north central and northeastern U.S. for more than half a century (Dennis et al., 1998; Eisen, Eisen and Beard, 2016; Gardner et al., 2020). It is hypothesized that the northern I. scapularis populations originated from migrant tick populations from the southern US that dispersed northward during the receding Pleistocene ice sheets (Humphrey, Caporale and Brisson, 2010; Frederick et al., 2023). The expansion of I. scapularis beginning in the middle of the last century was due to changes in land use as well as increases in the white-tailed deer (Odocoileus virginianus) populations due to conversion of agricultural land to forests and hunting restrictions on white-tailed deer (Barbour and Fish, 1993; Hamer et al., 2010; Ginsberg, Rulison, Miller, Pang, Arsnoe, Hickling, Ogden, LeBrun, et al., 2020). Ixodes scapularis is a habitat generalist (Labruna, 2014; Sonenshine, 2018a) and feeds on a wide range of different mammalian hosts which enable them to succeed in environments that humans inhabit. Thus, because of this rapid spread of I. scapularis in the Midwest and the Northeast, predicting the spatial risk of contracting I. scapularis and its associated diseases is important. In the Upper Midwest, the first populations of I. scapularis were found in western central 2 Wisconsin in the late 1960s (Gardner et al. 2020). In Michigan the first I. scapularis population was first detected in the Upper Peninsula, specifically in Menominee County in the 1980’s (Strand et al., 1992; Walker et al., 1994). Then, in 2002 the first I. scapularis population was detected in the southwestern corner of the Lower Peninsula of Michigan (Foster, 2004), after which I. scapularis has been expanding northward and eastward (Dennis et al., 1998; Eisen et al., 2016; Hamer et al. 2010; Lantos et al., 2017). Today, I. scapularis is detected in 63/83 counties of Michigan, some of which border Lake Huron on the east side of the state (Michigan Department of Health and Human Services, 2021, 2022). With the invasion of I. scapularis in Michigan, it is important to characterize the distribution of suitable habitats throughout Michigan and to better understand the abiotic and biotic factors that are important in predicting the future distribution of I. scapularis. The first predictive spatial model for I. scapularis in Michigan was based on applying a habitat suitability model developed from I. scapularis surveillance data in Wisconsin (Guerra et al., 2002), and it was from this model that Foster, and colleagues detected the first populations of I. scapularis in southwestern Michigan (Erik Scott Foster, 2004). Since then, there have been other models predicting the spatial distribution of I. scapularis within Michigan (Diuk-Wasser et al., 2010; Hahn et al., 2016; Burtis et al., 2022). Developing habitat suitability models specially in an area where I. scapularis is invading into, such as Michigan, is important because these models will help to target specific areas for future surveillance efforts. These models also have a public health benefit where we can inform the public as well as healthcare providers on the future risk of not just Lyme disease but other diseases that are caused by I. scapularis such as anaplasmosis. 3 Human anaplasmosis and the ecology of A. phagocytophilum Although Lyme disease is the leading tickborne disease in the U.S. and therefore is arguably the most important, anaplasmosis is the second leading tick-borne disease (Rosenberg et al., 2018). Human anaplasmosis was discovered in 1994 from western central Wisconsin and is caused by Anaplasma phagocytophilum, a gram-negative intracellular bacterium (Chen et al., 1994). Both A. phagocytophilum and B. burgdorferi share the same vector I. scapularis as well as major reservoir hosts such as the white-footed mouse (Peromyscus leucopus) (Donahue, Piesman and Spielman, 1987; Levin, Nicholson, Massung and Fish, 2002; Massung, Levin and Priestley, 2004; Stuen, Granquist and Silaghi, 2013). As with Lyme disease, the two major foci for human anaplasmosis are in the Upper Midwest and Northeast. Lyme disease and anaplasmosis share similar symptoms in humans, including flu-like symptoms with fever, headache, and malaise. Similar to Lyme disease, the peak onset of these symptoms is typically common during the late spring/early summer when nymphal I. scapularis are active (Chen et al., 1994). There is no transovarial transmission of A. phagocytophilum in its tick vector I. scapularis. The two I. scapularis life stages that are important in the transmission of A. phagocytophilum are the nymphs and the adults. Due to their relative smaller size and the activity period being in summer which overlaps with the time people are outside doing recreational activities, the nymphs possess relatively a greater disease risk than adults. Therefore, we see a lot of disease onset typically during summer periods. There is also a small bump in human disease cases during the fall and early spring. Because there is no transovarial transmission of A. phagocytophilum, the wildlife host community plays a major role in the enzootic cycle of A. phagocytophilum. A most comprehensive study that describes the ecology of A. phagocytophilum was conducted in the Northeast (Keesing 4 et al., 2012, 2014). No such study has been conducted in the Midwest which would help to understand the similarities and the differences between how A. phagocytophilum is maintained in the two regions. Thus, comparing how A. phagocytophilum maintained in the Midwest with that of the Northeast might help us uncover these epidemiological differences. This dissertation This dissertation examines two topics. One of the major components of this dissertation investigates an aspect of tick ecology while the other component investigates the ecology of a tickborne pathogen. The first topic focuses on studying the spread of I. scapularis within Michigan using habitat suitability models to describe where suitable habitats currently lie within Michigan. A region with a tick invasion is dynamic and over time habitats with endemic tick populations will expand but developing these models will help us in understanding the patterns of expansion over time. While there have been habitat suitability models developed for Michigan previously at a regional and national levels (Erik Scott Foster, 2004; Hahn et al., 2016; Burtis et al., 2022), I wanted to develop models based only on the comprehensive active surveillance data of I. scapularis available for Michigan. Furthermore, another goal in this study was to use two different methods of species distribution modeling to compare how similar or different they would be based on the different assumptions each method makes. Thus, this chapter will help to understand where suitable habitats are currently distributed in Michigan and will be the initial step towards predicting where I. scapularis may spread. With these initial habitat suitability models, we can make informed decisions on where our surveillance efforts should be targeted to and think where we should focus future tick prevention and control strategies. The second aspect of this dissertation focuses on the community ecology of A. 5 phagocytophilum within an I. scapularis endemic region in the Midwest. Lyme disease is the leading vector borne disease in the US; therefore, much of the ecology of its enzootic cycle has been studied but less research has been conducted on the ecology of other tickborne diseases. More thorough ecological studies of A. phagocytophilum have been conducted in the Northeast and to some extent in the West (where I. pacificus, the western blacklegged tick is the vector). Most studies on A. phagocytophilum in the Midwest have focused on tick ecology and less so on the host ecology with a few exceptions. Therefore, this chapter attempts to illuminate more the contribution of certain wildlife species on maintenance of A. phagocytophilum. With the increase in human anaplasmosis cases from the Midwest, it is important to learn more about the various ecological aspects of the maintenance of A. phagocytophilum in nature. Ixodes scapularis range expansion will result in the increase in disease incidence of multiple diseases. Thus, trying to understand the patterns of the invasion process of I. scapularis and how diseases spread by I. scapularis are maintained in nature will be helpful to guide prevention and control strategies in regions where the I. scapularis is endemic and in regions where I. scapularis is invading. Therefore, this dissertation hopes to bridge the knowledge gaps found in both ticks ecology and the ecology of A. phagocytophilum to better understand the spread of ticks and the maintenance of tickborne diseases. 6 BIBLIOGRAPHY Berrada, Z.L. and Telford, S.R. (2009) ‘Burden of Tick-borne Infections on American Companion Animals’, Topics in Companion Animal Medicine, 24(4), pp. 175–181. Available at: https://doi.org/10.1053/j.tcam.2009.06.005. Burtis, J.C. et al. (2022) ‘Predicting distributions of blacklegged ticks (Ixodes scapularis), Lyme disease spirochetes (Borrelia burgdorferi sensu stricto) and human Lyme disease cases in the eastern United States’, Ticks and Tick-borne Diseases, 13(5). Available at: https://doi.org/10.1016/j.ttbdis.2022.102000. Chen, S.M. et al. (1994) ‘Identification of a granulocytotropic Ehrlichia species as the etiologic agent of human disease.’, Journal of Clinical Microbiology, 32(3), pp. 589–595. Available at: https://doi.org/10.1128/jcm.32.3.589-595.1994. Dennis, D.T. et al. (1998) ‘Reported Distribution of Ixodes scapularis and Ixodes pacificus (Acari: Ixodidae) in the United States’, Journal of Medical Entomology, 35(5), pp. 629–638. Available at: https://doi.org/10.1093/jmedent/35.5.629. Diuk-Wasser, M.A. et al. (2010) ‘Field and climate-based model for predicting the density of host-seeking nymphal Ixodes scapularis, an important vector of tick-borne disease agents in the eastern United States’, Global Ecology and Biogeography, 19(4), pp. 504–514. Available at: https://doi.org/10.1111/j.1466-8238.2010.00526.x. Donahue, J.G., Piesman, J. and Spielman, A. (1987) ‘Reservoir competence of white-footed mice for Lyme disease spirochetes’, American Journal of Tropical Medicine and Hygiene, 36(1), pp. 92–96. Eisen, R.J. and Eisen, L. (2018) ‘The Blacklegged Tick, Ixodes scapularis: An Increasing Public Health Concern’, Trends in Parasitology. Elsevier Ltd, pp. 295–309. Available at: https://doi.org/10.1016/j.pt.2017.12.006. Eisen, R.J., Eisen, L. and Beard, C.B. (2016) ‘County-scale distribution of Ixodes scapularis and Ixodes pacificus (Acari: Ixodidae) in the continental United States’, Journal of Medical Entomology, 53(2), pp. 349–386. Available at: https://doi.org/10.1093/jme/tjv237. Eisen, R.J. and Paddock, C.D. (2021) ‘Tick and Tickborne Pathogen Surveillance as a Public Health Tool in the United States’, Journal of Medical Entomology, 58(4), pp. 1490–1502. Available at: https://doi.org/10.1093/jme/tjaa087. Fleshman, A.C. et al. (2021) ‘Reported County-Level Distribution of Lyme Disease Spirochetes, Borrelia burgdorferi sensu stricto and Borrelia mayonii (Spirochaetales: Spirochaetaceae), in Host-Seeking Ixodes scapularis and Ixodes pacificus Ticks (Acari: Ixodidae) in the Contiguous United States’, Journal of Medical Entomology. Edited by M. Diuk-Wasser, 58(3), pp. 1219– 1233. Available at: https://doi.org/10.1093/jme/tjaa283. 7 Foster, E.S. (2004) Ixodes scapularis (Acari:Ixodidae) and Borrelia burgdorferi in Southwest Michigan: Population ecology and verification of a geographic risk model. Michigan State University. Frederick, J.C. et al. (2023) ‘Phylogeography of the blacklegged tick (Ixodes scapularis) throughout the USA identifies candidate loci for differences in vectorial capacity’, Molecular Ecology, 32(12), pp. 3133–3149. Available at: https://doi.org/10.1111/mec.16921. Gardner, A.M. et al. (2020) ‘Landscape features predict the current and forecast the future geographic spread of Lyme disease: Landscape predicts Lyme disease spread’, Proceedings of the Royal Society B: Biological Sciences, 287(1941). Available at: https://doi.org/10.1098/rspb.2020.2278rspb20202278. Guerra, M. et al. (2002) ‘Predicting the risk of Lyme disease: Habitat suitability for Ixodes scapularis in the north central United States’, Emerging Infectious Diseases, 8(3), pp. 289–297. Available at: https://doi.org/10.3201/eid0803.010166. Hahn, M.B. et al. (2016) ‘Modeling the Geographic Distribution of Ixodes scapularis and Ixodes pacificus (Acari: Ixodidae) in the Contiguous United States’, Journal of Medical Entomology, 53(5), pp. 1176–1191. Available at: https://doi.org/10.1093/jme/tjw076. Hamer, S.A. et al. (2010) ‘Invasion of the lyme disease vector Ixodes scapularis: Implications for Borrelia burgdorferi endemicity’, EcoHealth, 7(1), pp. 47–63. Available at: https://doi.org/10.1007/s10393-010-0287-0. Hook, S.A. et al. (2022) ‘Economic Burden of Reported Lyme Disease in High-Incidence Areas, United States, 2014-2016’, Emerging Infectious Diseases, 28(6), pp. 1170–1179. Available at: https://doi.org/10.3201/eid2806.211335. Humphrey, P.T., Caporale, D.A. and Brisson, D. (2010) ‘Uncoordinated phylogeography of Borrelia burgdorferi and its tick vector, Ixodes scapularis’, Evolution, 64(9), pp. 2653–2663. Available at: https://doi.org/10.1111/j.1558-5646.2010.01001.x. Keesing, F. et al. (2012) ‘Reservoir Competence of Vertebrate Hosts for Anaplasma phagocytophilum’, Emerging Infectious Diseases, 18(12), pp. 10–13. Available at: https://doi.org/10.3201/eid1812.120919. Keesing, F. et al. (2014) ‘Prevalence of human-Active and variant 1 strains of the tick-borne pathogen Anaplasma phagocytophilum in hosts and forests of Eastern North America’, American Journal of Tropical Medicine and Hygiene, 91(2), pp. 302–309. Available at: https://doi.org/10.4269/ajtmh.13-0525 . Kugeler, K.J. et al. (2021) ‘Estimating the frequency of Lyme Disease diagnoses, United States, 2010-2018’, Emerging Infectious Diseases. Centers for Disease Control and Prevention (CDC), pp. 616–619. Available at: https://doi.org/10.3201/eid2702.202731. 8 Labruna, M.B. (2014) ‘Biology of Ticks edited by Daniel E. Sonenshine and R. Michael Roe’, The Quarterly Review of Biology, 89(4), pp. 402–403. Available at: https://doi.org/10.1086/678654. Lantos, P.M. et al. (2017) ‘Geographic Expansion of Lyme Disease in Michigan, 2000–2014’, Open Forum Infectious Diseases, 4(1), pp. 1–5. Available at: https://doi.org/10.1093/ofid/ofw269. Levin, M.L. et al. (2002) ‘Comparison of reservoir competence of medium sized mammals and Peromyscus leucopus for Anaplasma phagocytophilum in Connecticut’, Vector Borne and Zoonotic Disease, 2(3). Available at: https://doi.org/10.1089/ast.2007.0153. Massung, R.F., Levin, M.L. and Priestley, R.A. (2004) ‘Transmission Route Efficacy and Kinetics of Anaplasma phagocytophilum Infection in the White-Footed Mouse, Peromyscus leucopus’, Vector Borne and Zoonotic Diseases, 4(4), pp. 310–318. Available at: https://doi.org/10.1089/vbz.2004.4.310. Michigan Department of Health and Human Services (2021) Michigan Trends in Tickborne Disease, 2016-2020. Available at: https://www.michigan.gov//media/Project/Websites/emergingdiseases/Folder3/2021_Tickborne_ Disease_Summary_Report.pdf?rev=a77a79a5ca16467ebeef4a41c9272e55 (Accessed: 4 July 2023). Michigan Department of Health and Human Services (2022) Michigan Emerging and Zoonotic Disease Surveillance Summary 2021. Available at: https://www.michigan.gov/- /media/Project/Websites/emergingdiseases/EZID_Annual_Surveillance_Summary.pdf?rev=c41c 1aa053754235bc29f3d86c24b0c8 (Accessed: 4 July 2023). Miller, R.S., Farnsworth, M.L. and Malmberg, J.L. (2013) ‘Diseases at the livestock-wildlife interface: Status, challenges, and opportunities in the United States’, Preventive Veterinary Medicine, 110(2), pp. 119–132. Available at: https://doi.org/10.1016/j.prevetmed.2012.11.021. Pérez de Leon, A.A. et al. (2012) ‘Integrated strategy for sustainable cattle fever tick eradication in USA is required to mitigate the impact of global change’, Frontiers in Physiology, 3 JUN. Available at: https://doi.org/10.3389/fphys.2012.00195. Rodino, K.G., Theel, E.S. and Pritt, B.S. (2020) ‘Tick-Borne Diseases in the United States’, Clinical chemistry. NLM (Medline), pp. 537–548. Available at: https://doi.org/10.1093/clinchem/hvaa040. Rosenberg, R. et al. (2018) ‘Vital Signs: Trends in Reported Vectorborne Disease Cases — United States and Territories, 2004–2016’, MMWR. Morbidity and Mortality Weekly Report, 67(17), pp. 496–501. Available at: https://doi.org/10.15585/mmwr.mm6717e1. Sonenshine, D.E. (2018) ‘Range expansion of tick disease vectors in north america: Implications for spread of tick-borne disease’, International Journal of Environmental Research and Public 9 Health. MDPI. Available at: https://doi.org/10.3390/ijerph15030478 Strand, M.R., Walker, E.D. and Merritt, R.W. (1992) ‘Field studies on Ixodes dammini in the Upper Peninsula of Michigan’, Vector Control Bulletin of North Central States, 1, pp. 11–18. Stuen, S., Granquist, E.G. and Silaghi, C. (2013) ‘Anaplasma phagocytophilum--a widespread multi-host pathogen with highly adaptive strategies.’, Frontiers in cellular and infection microbiology, 3(July), p. 31. Available at: https://doi.org/10.3389/fcimb.2013.00031. Tsao, J.I. et al. (2021) ‘The Contribution of Wildlife Hosts to the Rise of Ticks and Tick-Borne Diseases in North America’, Journal of Medical Entomology, 58(4), pp. 1565–1587. Available at: https://doi.org/10.1093/jme/tjab047. Walker, E.D. et al. (1994) ‘Prevalence of Borrelia burgdorferi in Host-Seeking Ticks (Acari: Ixodidae) from a Lyme Disease Endemic Area in Northern Michigan’, Journal of Medical Entomology, 31(4), pp. 524–528. Available at: https://doi.org/10.1093/jmedent/31.4.524. Wisely, S.M. and Glass, G.E. (2019) ‘Advancing the science of tick and tick-borne disease surveillance in the United States’, Insects. MDPI AG. Available at: https://doi.org/10.3390/insects10100361. World Health Organization (2017) Global vector control response 2017 - 2030. Geneva. 10 CHAPTER 2: ANAPLASMA PHAGOCYTOPHILUM: A REVIEW ABSTRACT Anaplasma phagocytophilum is the causative agent of human granulocytic anaplasmosis in the US and Europe. It is an emerging vector borne disease spread by ticks in the Ixodes ricinus complex throughout the world within the northern hemisphere. Clinical symptoms of A. phagocytophilum are flu-like and symptoms typically are seen in summer and fall, coinciding with the activity periods of the vector stages. It is an intracellular pathogen, generally infecting neutrophils, with intricate mechanisms to infect a host cell and evade the host immune system. Anaplasma phagocytophilum is known to infect several vertebrate species, where several serve as important reservoir hosts. Furthermore, Anaplasma phagocytophilum seems to show host associations, such that some strains are known to be exclusively circulating within specific species. In this literature review here, I provide a basic understanding of the pathogenesis, ecology, and epidemiology of A. phagocytophilum. Keywords: Anaplasma phagocytophilum, pathogenesis, ecology, epidemiology, host, vectors 11 INTRODUCTION Anaplasma phagocytophilum is an emerging vector borne pathogen. It is known to cause human granulocytic anaplasmosis, tick-borne fever in ruminants, equine granulocytic anaplasmosis in horses, febrile fever in cats, and canine anaplasmosis in dogs (Rar and Golovljova, 2011). In humans, typical disease symptoms include fever, malaise, myalgia, thrombocytopenia, and leukopenia (Bakken and Dumler, 2015). Symptoms in dogs and cats are very similar, including lethargy, fever, and anorexia (Little, 2010). Most human patients have mild clinical symptoms and do not require hospitalizations; the case fatality rate for A. phagocytophilum is relatively low and is about 0.6% (Dumler, 1997). Distribution and epidemiology of A. phagocytophilum Anaplasma phagocytophilum has been detected in parts of Europe (e.g., Norway, United Kingdom, Germany, Slovakia, and Switzerland), Asia (e.g., Russia, China, and Korea), and North America, where the major foci are found in Upper Midwest and the Northeast US and where a third much smaller foci occur in western US (Stuen, 2007; Stuen, Granquist and Silaghi, 2013). In the US, it was first discovered in patients from northern Minnesota and Wisconsin in 1992, when one who presented with severe flu-like symptoms two weeks after a tick bite died (Chen et al., 1994). Since then, the number of human granulocytic anaplasmosis cases has increased (Biggs et al., 2016) and has become increasingly important for public health. The incidence rates of human cases have increased from 2.0 cases per million persons from 2000-to 2007 to 6.3 cases per million cases from 2008-to 2012 (Rar and Golovljova, 2011; Dahlgren et al., 2015). According to the CDC, human granulocytic anaplasmosis is the second most reported vector-borne disease, after Lyme disease (Biggs et al., 2016; Baker et al., 2020; Elias et al., 2020), and areas with high Lyme disease risk also have a high risk of human granulocytic anaplasmosis. 12 Seroepidemiological data from people suggest that in endemic regions 15% - 36% of the population seem to have been infected with human granulocytic anaplasmosis (Bakken and Dumler, 2015). Human cases are relatively higher among males who are older than 40 years old, while people with a compromised immune system are more susceptible (Jin et al., 2012; Bakken and Dumler, 2015; Biggs et al., 2016; Dumic et al., 2022). Human disease incidence in Europe and Asia seems to be less common compared to that in the US (Dumler et al., 2005; Stuen, 2007; Stuen, Granquist and Silaghi, 2013; Dumic et al., 2022a), which may be due to the low abundance of the A. phagocytophilum strains that infects and causes disease in humans. In dogs, canine anaplasmosis was first detected from a dog in California and since then is found throughout the US in regions where I. scapularis is endemic and invading; considering serological data, up to about 40% of dogs are found to be seropositive for canine granulocytic anaplasmosis (Beall et al., 2008; Bowman et al., 2009; Carrade et al., 2009; Qurollo et al., 2014; Khatat et al., 2021). Pathogenesis of infection Anaplasma phagocytophilum is a gram-negative intracellular pathogen of about 0.4 – 1.3 mm in size (Severo et al., 2012). It can infect mammalian cells of hematopoietic origin; it primarily infects neutrophils, and to a lesser extent, monocytes, macrophages, red blood cells, platelets, and endothelial cells (Rikihisa, 2011). The pathogenesis of A. phagocytophilum is poorly understood. Anaplasma phagocytophilum is polymorphic bacteria where during infection of a host cell, a dense-cored form binds to the cell, and then once internalized in the host neutrophil; the early forms develop into round reticular structures residing within an early endosome, acquire nutrients through binary fission, and develop into distinct membrane-bound intracytoplasmic bacterial aggregates called morulae, which later on become small dense structures (Chen et al., 1994; Dumler, 1997; Webster et al., 1998; McQuiston et al., 1999; Dumler et al., 2005; Rikihisa, 2010; 13 Bakken and Dumler, 2015). These morulae are apparent in microscopic slides of stained blood smears and tissue samples prepared from infected patient tissues (Chen et al., 1994; Dumler et al., 2001, 2005). Anaplasma phagocytophilum has a relatively small genome size of about 1.47 – 148 Mb with the absence of plasmids (Dunning Hotopp et al., 2006; Barbet et al., 2013). Anaplasma phagocytophilum lacks the genes needed for the synthesis of lipopolysaccharides and peptidoglycan, and consequently, the outer membrane lacks the peptidoglycan and lipopolysaccharide layers (Lin and Rikihisa, 2003b). Due to this lack of peptidoglycan and lipopolysaccharide, A. phagocytophilum cells incorporate host cell cholesterol to survive (Lin and Rikihisa, 2003b). During the initial infection process, A. phagocytophilum binds to the P-selectin binding domain of the human neutrophils through the P-selectin glycoprotein ligand (PSGL-1) on the A. phagocytophilum cell (Herron et al., 2000). The dense-core form of A. phagocytophilum facilitates the adhesion to the host cell by recognizing the human PSGL-1 (Troese and Carlyon, 2009). Most human neutrophil cells are enriched with P-selectin binding domain cell surface receptors, making it easy for A. phagocytophilum to attach and enter human neutrophils. When infecting mouse neutrophils, A. phagocytophilum P-selectin glycoprotein ligand is not required; instead, fucosyl transferases are required (Rikihisa, 2011). Several surface proteins, such as major surface proteins Msp2, Asp55, and Asp62, are required for A. phagocytophilum to bind and infect the host cells (Wang, Kikuchi and Rikihisa, 2006; Ge and Rikihisa, 2007). Entry into the host cell is mediated through lipid rafts known as caveolae (Lin and Rikihisa, 2003a). Through lipid caveolae, A. phagocytophilum cells are directed into inclusions within the host cell. These A. phagocytophilum inclusions do not have characteristic features of an early cellular endosome such as the early endosome marker Rab5, early endosome antigen 1 (EEA1), the vacuolar (H +) ATPase 14 (Webster et al., 1998; Mott, Barnewall and Rikihisa, 1999; Yoshiie et al., 2000).Once inside these inclusions, replication of A. phagocytophilum occurs (Rikihisa, 2011). The replicative form of A. phagocytophilum is the reticulate cellular form, and replication occurs by binary fission (Troese and Carlyon, 2009). During the replication process, which takes about 24 hours, the membrane of parasitophorous vacuoles (morulae) increases in size to accommodate the multiplying reticulate form of A. phagocytophilum cells (Troese and Carlyon, 2009). After 24 hours of replication, the reticulate forms start to condense into the dense-core form and are ready to burst out of the host cell and infect new host cells (Troese and Carlyon, 2009). Neutrophils are an essential part of the innate immune system known for their ability to phagocytose invading pathogens and lysing them through the formation of lysosomes. Neutrophils generally have a short lifespan, and a short half-life of about 7 hours (Tak et al., 2013). A characteristic feature of a neutrophil is the ability of spontaneous apoptosis to enable cell turnover and homeostasis in the host system. Anaplasma phagocytophilum can inhibit a neutrophil's spontaneous apoptosis ability, thereby increasing the lifespan of the neutrophil. This inhibition process provides the A. phagocytophilum bacterium sufficient time to develop and replicate within the infected neutrophil (Scaife et al., 2003; Borjesson et al., 2005; Ge et al., 2005; Tak et al., 2013). Anaplasma phagocytophilum inhibits spontaneous apoptosis in neutrophils in several ways by upregulating anti-apoptotic genes, by maintaining the membrane potential of mitochondrial membrane within infected neutrophils, and by inhibiting the caspase3 activation pathway required for spontaneous cell apoptosis (Ge et al., 2005). Autophagy is a natural process of cellular degradation mediated by lysosomes. In a bacterial infection, autophagy is vital to a cell because it helps clear out intracellular infections and will enable the infected cell to differentiate between self and pathogen antigens to be presented to activate the innate and active immune responses (Amano, 15 Nakagawa and Yoshimori, 2006). Once A. phagocytophilum infects the host cell, it subverts the autophagy response mechanism of the host neutrophil (Niu, Yamaguchi and Rikihisa, 2008). After invading a host neutrophil, A. phagocytophilum inclusions show characteristic features of an early autophagosome, including being enveloped by a double-lipid bilayer membrane and localization of essential components of cellular autophagosomes such as microtubule-associated proteins and beclin-1 to the A. phagocytophilum inclusions (Niu, Yamaguchi and Rikihisa, 2008). Although A. phagocytophilum inclusions shows autophagosome characteristics, it does not mature into a late autophagosome and does not fuse with an autolysosome (Niu, Yamaguchi and Rikihisa, 2008). Thus, A. phagocytophilum can develop and replicate safely within these autophagosomes without being lysed. During the infection of A. phagocytophilum, patients develop a humoral immunity and launch pro-inflammatory cytokine responses (Dumler et al., 2000). Typically, patients infected with anaplasmosis are treated with doxycycline, and many patients seem to develop high titers of antibodies again anaplasmosis which lasts up 12 – 18 months (Bakken and Dumler, 2015). Tick vectors Anaplasma phagocytophilum is transmitted by hard ticks in the Ixodes ricinus complex. In the eastern US, it is transmitted by the blacklegged tick, I. scapularis; in the western US, it is transmitted by the western blacklegged tick, I. pacificus; in Europe, the primary vector is the sheep tick or the castor bean tick, I. ricinus; and in Asia and Russia, the primary vector is I. persulcatus (Woldehiwet, 2010; Stuen, Granquist and Silaghi, 2013; Dugat et al., 2015). Anaplasma phagocytophilum has been detected by PCR in tick species such as Dermacentor albipictus (the winter tick), Haemaphysalis leporispalustris (the rabbit tick) and H. longicornis (the Asian long- horned tick) (Goethert and Telford, 2003; Baldridge et al., 2009; Price et al., 2022), while the vector competence for D. albipictus and H. leporispalustris has not been shown (Stuen, Granquist 16 and Silaghi, 2013) whereas H. longicornis was shown not to be a competent vector in transmitting A. phagocytophilum (Levin et al., 2021). Ticks in the Ixodes genus are three-host feeders, and at each life stage, the tick blood feeds on a new host except for adult male ticks, which do not feed. Anaplasma phagocytophilum is not transmitted vertically in Ixodes species; thus, only the nymphal and adult (female) stages can transmit the bacterium to mammalian and avian hosts. Therefore, for an Ixodes spp. tick to acquire A. phagocytophilum, it must feed on an infected host. In Dermacentor albipictus ticks, the transovarial transmission of A. phagocytophilum under laboratory conditions has been observed (Baldridge et al., 2009). In addition, several studies have shown that Ixodes spp. ticks could acquire A. phagocytophilum through co-feeding, where transmission of the pathogen can occur from an infected tick to an uninfected tick feeding in proximity during simultaneous feeding on an uninfected host (Levin and Fish, 2000a; Ogden et al., 2002). When an Ixodes spp. tick feeds on an infected host, A. phagocytophilum will reach the ticks midgut through the blood meal. Within a tick, A. phagocytophilum survives in the midgut cells and later migrates into the salivary gland cells, and this process will take about 24 hours of feeding on the infected host (Hodzic et al., 2001; Severo et al., 2012). The migration of A. phagocytophilum from the midgut to salivary glands is mediated through the hemolymph of the tick. A tick salivary protein called P11 facilitates the migration of A. phagocytophilum by enabling infection of tick hemocytes (Liu et al., 2011). The infection of hemocytes by A. phagocytophilum is vital to infecting the salivary glands. Anaplasma phagocytophilum produces inclusions and has been shown to reside within the secretory salivary acini (Telford et al., 1996). Once an infected tick starts blood-feeding on a mammalian host, the feeding will activate the migration of A. 17 phagocytophilum from the midgut into the salivary glands, and within the salivary glands A. phagocytophilum will start replicating (Hodzic et al., 2001). During the infection within the tick, A. phagocytophilum manipulates the secretion of several proteins within the tick to survive and be transmitted into a viable host. For example, A. phagocytophilum induces the expression of a salivary protein, salp16, within I. scapularis ticks (Sukumaran et al., 2006). Sukumaran et al., 2006 have shown that during early infection of the salivary glands, A. phagocytophilum requires salp16 protein. Another protein that is upregulated in I. scapularis upon the infection of A. phagocytophilum is the tick antifreeze protein (Neelakanta et al., 2010), which may enable A. phagocytophilum-infected I. scapularis to survive better during winter compared to the non-infected I. scapularis ticks. Therefore I. scapularis ticks may benefit from this interaction with A. phagocytophilum, although evidence that A. phagocytophilum- infected ticks have higher over-wintering survivorship compared to non-infected ticks in nature has yet to be demonstrated. Several experiments have shown that transmission of A. phagocytophilum from an infected I. scapularis tick to a host occurs between 24 and 48 hours after the tick attaches to the host (Hodzic et al., 1998; des Vignes et al., 2001; Levin, Troughton and Loftis, 2021). The transmission from an infected tick to a naïve host is dose-dependent, and in laboratory mice (Mus musculus) a median dose of 104 – 105 of morulae are required to infect a mouse (Hodzic et al., 1998). Therefore, the initial quantity of A. phagocytophilum within the tick may be insufficient to infect a host and A. phagocytophilum must replicate within the tick to achieve the infectious dose. The number of A. phagocytophilum copies within I. scapularis has been shown to increase over a 72-hour tick feeding period allowing a sufficient infectious dose (Levin, Troughton, and Loftis, 2021). Therefore, the longer an infected tick is attached to a naive host, the greater the potential of the 18 host getting infected with A. phagocytophilum; for reference, I. scapularis nymphs typically complete their bloodmeal in about 4 days on a laboratory rabbit (Troughton and Levin, 2007). Nymphal ticks are the most crucial stage in transmitting the pathogen to humans because they are the first life stage capable of infecting a host, numerous, difficult to detect due to their small size, and because of the nymphal activity peaks in late spring/early summer coinciding when humans are highly active outdoors (Biggs et al., 2016; Murphy et al., 2017). Adult female ticks also can carry A. phagocytophilum and infect humans. Generally, since adult female ticks are larger than nymphs and have a reddish abdomen, the probability is greater that they will be discovered before taking a blood meal or before feeding long enough to transmit an infectious dose of A. phagocytophilum compared to nymphal ticks (Falco, Richard C., Durland, 1988). As discussed earlier, it takes at least 24 – 48 hours to transmit A. phagocytophilum from an infected I. scapularis to a host. Due to the relatively shorter transmission time compared with that of the Lyme disease pathogen, which requires at least 36 hours of feeding, adult I. scapularis may also play a larger role in human disease transmission for anaplasmosis compared with Lyme disease. The presence of a small peak of human granulocytic anaplasmosis cases seen during the fall when adult blacklegged ticks are active but nymphs are not, and the absence of such a peak for Lyme disease, supports the hypothesis that adult I. scapularis may play a relatively larger epidemiological role. Reservoir hosts of A. phagocytophilum Anaplasma phagocytophilum utilizes similar reservoir hosts as Borrelia burgdorferi, the Lyme disease bacterium (Keesing et al., 2012, 2014; Stuen, Granquist and Silaghi, 2013; Foley et al., 2016; Stephenson and Foley, 2016). Similar to B. burgdorferi, the primary reservoir for A. phagocytophilum in the northeastern and northern midwestern US is the white-footed mouse 19 (Peromyscus leucopus) (Telford et al., 1996; Magnarelli et al., 1997; Walls et al., 1997; Levin, Nicholson, Massung, Sumner, et al., 2002), which is mainly found in woodland habitats (Musser, 1969). Once infected, white-footed mice can launch a strong humoral immune response within 7 – 14 days after exposure (Levin and Fish, 2000b).From laboratory xenodiagnoses experiments, infected laboratory mice (M. musculus) can transmit A. phagocytophilum to naïve larvae for up to 9 weeks (Levin and Ross, 2004), although the peak transmission period (i.e., infecting the greatest proportion of naïve larvae) occurs within the first 2-3 weeks, after which transmission efficiency decreases. Several studies have shown that P. leucopus have very high larval burdens compared to other small mammal species (Keesing et al., 2009; Hersh et al., 2014), which is important in maintenance of the enzootic cycle of A. phagocytophilum. Several other small mammals in the northeastern and upper midwestern US are also known to be competent reservoir hosts for A. phagocytophilum. These include the eastern chipmunk (Tamias striatus), northern short-tailed shrew (Blarina brevicauda), red squirrel (Tamiasciurus hudsonicus), flying squirrel (Glaucomys volans), gray squirrel (Sciurus carolinensis), masked shrew (Sorex cinereus), meadow jumping mouse (Zapus hudsonicus), red-backed vole (Clethrionomys gapperi), and meadow vole (Microtus pennsylvanicus) (Walls et al., 1997; Stafford et al., 1999; Levin, Nicholson, Massung and Fish, 2002; Johnson et al., 2011; Keesing et al., 2012, 2014). In addition to these small mammals, raccoons (Procyon lotor) and cottontail rabbits (Sylvilagus floridanus) can become infected with A. phagocytophilum (Levin, Nicholson, Massung and Fish, 2002; Goethert and Telford, 2003). Procyon lotor appears to have lower reservoir competence for A. phagocytophilum (Keesing et al., 2012) compared to mice; no reservoir competence studies for S. floridanus have not been conducted. In the US, many passerine bird species serve as hosts for I. scapularis larvae and nymphal 20 ticks. For example, the gray catbird (Dumetella carolinensis), Swainson's thrush (Catharus ustulatus), American robin (Turdus migratorius), wood thrush (Hylocichla mustelina), eastern towhee (Pipilo erythrophthalmus), brown thrasher (Toxostoma rufum), Carolina wren (Thryothorus ludovicianus), Northern cardinal (Cardinalis cardinalis), ovenbird (Seiurus aurocapilla), veery (Catharus fuscescens), house wren (Troglodytes aedon), chipping sparrow (Spizella passerina), indigo bunting (Passerina cyanea), and common yellowthroat (Geothlypis trichas) are some of the common passerine bird species that are commonly infested with nymphal and larval life stages of I. scapularis (Anderson et al., 1986; Stafford, Bladen and Magnarelli, 1995; Nicholls and Callister, 1996; Smith et al., 1996; Scharf, 2004; Hamer, Goldberg, et al., 2012; Scott, Anderson and Durden, 2012). Although birds can be infested with I. scapularis, the ability of these bird species to act as competent reservoir hosts for A. phagocytophilum seems to be relatively low, or A. phagocytophilum has not been detected in these species (Daniels et al., 2002; Ogden et al., 2008; Hamer, Goldberg, et al., 2012; Johnston et al., 2013; Dingler et al., 2014; Dumas et al., 2022). For example, in a study in upstate New York in the Northeast, the reservoir competence of A. phagocytophilum in four bird species - veery, gray catbird, wood thrush, and American robin - ranged from 2% - 10% (Keesing et al., 2012) which is relatively low compared to some of the small mammal species described previously which was greater than 10%. In California, the enzootic cycle of A. phagocytophilum is slightly different. The wildlife hosts for I. pacificus are somewhat different than that in the northern eastern US in that the immature ticks commonly feed on lizards, birds, and small mammals, while adult I. pacificus feed on deer, dogs, coyotes, bears, bobcats, and numerous other hosts (Furman and Loomis, 1984). In the western USA, the dusky-footed woodrat (Neotoma fuscipes) plays an important role in the enzootic cycle of A. phagocytophilum (Nicholson et al., 1999). Other mammalian hosts known to 21 be infected with A. phagocytophilum in the western US include redwood chipmunk (Tamias ochrogenys), brush mouse (Peromyscus boylii), pinyon mouse (Peromyscus truei), western harvest mouse (Rheithrodontomys megalotis), western grey squirrel (Sciurus griseus), American black bear (Ursus americanus), and gray fox (Urocyon cinereoargenteus) (Foley et al., 2004; Drazenovich, Foley and Brown, 2006; Foley, Clueit and Brown, 2008; J. E. Foley et al., 2008; Nieto and Foley, 2008; Gabriel et al., 2009; Foley and Nieto, 2011). Lizards and reptiles in California such as northern alligator lizard (Elgaria coereleus), sagebrush lizard (Sceloporus graciosus), western fence lizard (Sceloporus occidentalis), Pacific gopher snake (Pituophis catenifer), and common garter snake (Thamnophis sirtalis) are also shown to become infected with A. phagocytophilum. However, their reservoir competence is very low (Nieto et al., 2009). In Europe, the reservoir host composition responsible for maintaining A. phagocytophilum differs from the United States. In Europe, small mammals, particularly rodents, may not play a significant role in maintaining A. phagocytophilum in nature. Several studies have shown rodents species such as a yellow-necked mouse (Apodemus flavicollis), wood mouse (Apodemus sylvaticus), black-striped field mouse (Apodemus agrarius), and several different vole species such as the bank vole (Myodes glareolus), common vole (Microtus arvalis), field vole (Microtus agrestis) and root vole (Microtus oeconomus) to be infected with A. phagocytophilum at a very low level, but their reservoir competence has not been studied (Jorge S. Liz et al., 2000; Kevin J Bown et al., 2003; Hulínská et al., 2004; Grzeszczuk et al., 2006; Smetanová, Schwarzová and Kocianová, 2006; Barandika et al., 2007; Marumoto et al., 2007; Silaghi, Woll, et al., 2012; Majazki et al., 2013). Few shrew species, such as the common shrew (Sorex araneus) and the greater, white-tooted shrew (Crocidura russula), are also known to be infected with A. phagocytophilum in the UK, Switzerland, and Spain (Ogden et al., 1998; Jorge S. Liz et al., 2000; 22 Kevin J Bown et al., 2003; Barandika et al., 2007; Bray et al., 2007). A few recent studies in the UK, Germany, and Romania have shown that the European hedgehog (Erinaceus europaeus) and the Northern, white-breasted hedgehog (Erinaceus roumanicus) have high infection prevalence for A. phagocytophilum and are hypothesized to be competent reservoir hosts for A. phagocytophilum (Silaghi, Skuballa, et al., 2012; Dumitrache et al., 2013; Földvári et al., 2014). In Europe, wild ruminants may play a greater role in the enzootic cycle of A. phagocytophilum, where roe deer (Capreolus capreolus), red deer (Cervus elaphus), fallow deer (Dama dama), sika deer (Cervus Nippon) are reported to have relatively high A. phagocytophilum infection prevalence in the UK, Denmark, Poland, Slovakia, Czech Republic, Germany, Austria, Switzerland, and in Italy (Oporto et al., 2003; Beninati et al., 2006; Carvalho et al., 2008; Veronesi et al., 2011; Overzier et al., 2013; Stuen et al., 2013; Kauffmann et al., 2017; Stigum et al., 2019; Remesar et al., 2020; Silaghi et al., 2020). In Europe, I. ricinus is known to infest several bird species. Many bird species throughout Europe are not competent reservoir hosts for A. phagocytophilum due to their inability to become infected (Bjöersdorff et al., 2001; Skotarczak et al., 2006; Franke et al., 2010; Hildebrandt et al., 2010). Several studies have shown the common blackbird (Turdus merula) to be infected with A. phagocytophilum and to carry infected larvae indicating that they could be a potential competent reservoir host of A. phagocytophilum (Fuente et al., 2005; Skotarczak et al., 2006; Paulauskas, Radzijevskaja and Rosef, 2009; Palomar et al., 2012; Jahfari et al., 2014; Mărcuţan et al., 2014). Apart from the common blackbird, these studies have found the common chaffinch (Fringilla coelebs), red wing (Turdus iliacus), song thrush (Turdus philomelos), house sparrow (Passer domesticus), Spanish sparrow (Passer hispaniolensis), rock bunting (Emberiza cia), woodchat shrike (Lanius senator), magpie (Pica pica) and long-tailed tit (Aegithalos caudatus) either to be 23 infected with A. phagocytophilum or harbor infected larvae, indicating transmission of A. phagocytophilum to larvae (Fuente et al., 2005; Paulauskas, Radzijevskaja and Rosef, 2009; Hornok et al., 2014; Mărcuţan et al., 2014). Although birds in Europe and North America may not play a prominent role in the enzootic cycle of A. phagocytophilum, they may contribute to the dispersal of infected ticks. Several of these bird species were shown to have infected I. ricinus nymphs indicating the potential threat of birds dispersing infected ticks into unexplored areas where previous A. phagocytophilum infections were unknown. Further investigation into the common blackbird's role in the enzootic cycle of A. phagocytophilum should be studied as it is a relatively common bird species in Europe. Genetic diversity of A. phagocytophilum Europe has been at the forefront of strain diversity studies for A. phagocytophilum, in part perhaps of the importance to livestock. Anaplasmosis in Europe predominantly is a disease of livestock. The strains circulating in Europe mainly cause tick-borne fever in ruminants, especially in sheep and cattle (Ladbury et al., 2008; Beugnet and Marié, 2009; Atif, 2015). Based on the conserved 16S rRNA gene in the US, two A. phagocytophilum strains are commonly found, including a human pathogenic strain (Ap-ha), which causes disease in humans, and a deer variant strain (Ap-v1) that is not known to cause disease in humans (Massung et al., 1998, 2002). Although both strains have been detected in multiple wildlife species (Keesing et al., 2014) the primary reservoir host for Ap-ha is the white-footed mouse (P. leucopus), which is only weakly competent for Ap-v1, while white-tailed deer serve as the primary reservoir host for Ap-v1 and are incompetent for Ap-ha (Massung et al., 1998, 2003, 2005; Stafford et al., 1999). Although Ap-v1 (and closely related strains) was first discovered in the north central US (Massung et al., 1998; Michalski et al., 2006) very little work has been conducted to investigate the ecology, 24 including host associations, of different strains. White-footed mice, chipmunks, and raccoons showed lower reservoir competence to Ap-v1 than Ap-ha (Yabsley et al., 2008; Keesing et al., 2014). Considering birds, catbirds, veeries, and wood thrush showed a low reservoir competence for Ap-ha, while American robins had low reservoir competence for both Ap-v1 and Ap-ha (Keesing et al., 2012). Of note, Ap-ha is rarely detected in Europe although cases of human granulocytic anaplasmosis have (Rar, Tkachev and Tikunova, 2021; Dumic et al., 2022a) been reported (Atif, 2015; Lagler et al., 2017; Tsiodras et al., 2017; Dumic et al., 2022a), and there have not been enough studies conducted in Asia to understand its prevalence there. In California (US), researchers have explored the genetic diversity of A. phagocytophilum more thoroughly than in the northeastern and north central US, using multiple genetic markers. Several early studies differentiating genetic strains of A. phagocytophilum used 16S rRNA. In Norway, several variants of 16S rRNA were found in sheep with distinct clinical manifestations (Stuen et al., 2002, 2003). In addition, early European studies used the 16S rRNA gene to distinguish variants of A. phagocytophilum circulating in red deer and roe deer (Zeman and Pecha, 2008). However, different studies have concluded that the level of variation within the 16S rRNA gene may not be enough to delineate among strain types in Europe, which led to researchers investigating other genes(Bown et al., 2009; Scharf et al., 2011; Silaghi, Liebisch and Pfister, 2011), including those encoding the ankyrin protein (ankA gene), the groESL operon (groES gene, and groEL gene), and the major surface protein 2 (msp2 gene). Researchers in Europe and Russia further have developed a multilocus sequence typing (MLST) system to investigate A. phagocytophilum genetic diversity (Huhn et al., 2014; Mukhacheva, Shaikhova and Kovalev, 2019; Mukhacheva et al., 2020). 25 The gene that encodes ankyrin protein (AnkA gene) has been one commonly used gene to differentiate A. phagocytophilum strains mainly in Europe. Ankyrin proteins are essential in different protein-protein interactions and modulate gene transcription within the host cells (Caturegli et al., 2000; Loewenich et al., 2003; Park et al., 2004; Ijdo, Carlson and Kennedy, 2007; Scharf et al., 2011; Majazki et al., 2013; Mukhacheva et al., 2020). The ankA gene target can differentiate between five A. phagocytophilum variant clusters from different host species (Scharf et al., 2011; Dugat et al., 2015). The first gene cluster was including sequences taken from humans, dogs, cats, horses, and a few ruminants (Rar and Golovljova, 2011). This ankA gene cluster branched into two variants, grouped geographically (i.e., European and versus American regions origins) (Rar and Golovljova, 2011). The rest of the gene clusters were all from different host species in Europe, where the second gene cluster comprised variants from roe deer and red deer; the third gene cluster comprised variants from cattle, sheep, roe deer, and red deer; the fourth gene cluster comprised variants from roe deer; and the fifth gene cluster comprised variants from rodents (Rar and Golovljova, 2011; Scharf et al., 2011). Therefore, ankA gene cluster may show varying degrees of host association. The groESL operon in A. phagocytophilum covers the region of two encoding genes, groES, and groEL, which are essential in heat shock protein production. Much of the groESL genotyping has been conducted in Europe and Russia. Four major groESL clusters of A. phagocytophilum have been delineated (Jahfari et al., 2014) and are referred to as ecotypes that appear to be host specific. The most common of these gene clusters is ecotype I, comprising isolates from humans, dogs, cattle, horses, hedgehogs, red deer, sheep, and mouflons. Ecotype II isolates are primarily from roe deer; ecotype III isolates are from rodents and questing I. persulcatus; and ecotype IV isolates are often associated with birds- mainly the common blackbird 26 (Jahfari et al., 2014). Since both AnkA and groESL genes show similar clustering patterns, there is more evidence to support the hypothesis that A. phagocytophilum strains show host tropisms. In both gene targets, gene sequences originating from America cluster separately from the European sequences, indicating distinguishing variants of A. phagocytophilum circulating within different regions. A major surface protein gene of A. phagocytophilum (msp2) also shows a regional clustering pattern, where the American variants cluster separately from the European variants (Morissette et al., 2009; Silaghi, Liebisch and Pfister, 2011). This clear regional clustering pattern may be due to differences in host species and host species composition between the two regions as well as divergence due to geographic separation. In California, a genetic variant based on the msp2 gene of A. phagocytophilum cycles exclusively within dusky-footed woodrats (Foley et al., 2008). This variant does not infect horses or dogs, is distinct from Ap-ha and Ap-v1 variants and from the ankA gene variants (Trost et al., 2018). Dusky-footed woodrats frequently show high infection prevalence and seropositivity in California (Nicholson et al., 1999; J. Foley et al., 2008). Recent advancement in strain typing A. phagocytophilum uses multi-locus sequence typing (MLST) techniques. This approach uses seven housekeeping genes to determine the sequence types. However, there have been only four studies done using MLST for A. phagocytophilum, and all are based on samples collected in Europe and Russia. Overall, using MLST, it was found that the sequence typing showed similar patterns with the 16S rRNA and ankaA loci (Chastagner et al., 2014; Huhn et al., 2014; Mukhacheva, Shaikhova and Kovalev, 2019; Mukhacheva et al., 2020). The future in strain typing of A. phagocytophilum should be based on whole-genome sequencing. However, only a few studies have sequenced the genome of A. phagocytophilum (Dunning Hotopp et al., 2006; Barbet et al., 2013; Dugat et al., 2015). Thus, there have been no 27 publications, to our knowledge, using whole genome sequencing to investigate genetic diversity and host association of A. phagocytophilum, indicating the need to sequence more variants to identify different strains circulating within different regions and within different host species. Conclusion There are many gaps in knowledge in the ecology of A. phagocytophilum especially in the Upper Midwest, where it is endemic. There are also many gaps in knowledge in the strain diversity of A. phagocytophilum in the Upper Midwest and the Northeast of the US. With the improvements seen in whole genome sequencing techniques in the last few years, looking at strain diversity within endemic regions of the eastern USA would be important in understand how A. phagocytophilum strains are maintained in nature and their importance to human disease risk. Furthermore, not a lot is known about the pathogenesis of A. phagocytophilum as its transmitted from the tick to the host and vice versa. With the continued range expansion of I. scapularis in the eastern U.S., further investigating the ecology, strain diversity and pathogenesis of A. phagocytophilum will be crucial in reducing disease risk. 28 BIBLIOGRAPHY Amano, A., Nakagawa, I. and Yoshimori, T. (2006) ‘Autophagy in innate immunity against intracellular bacteria’, Journal of Biochemistry, 140(2), pp. 161–166. Available at: https://doi.org/10.1093/jb/mvj162. Anderson, J.F. et al. (1986) ‘Involvement of birds in the epidemiology of the Lyme disease agent Borrelia burgdorferi’, Infection and Immunity, 51(2), pp. 394–396. Available at: https://doi.org/10.1128/iai.51.2.394-396.1986. Atif, F.A. (2015a) ‘Anaplasma marginale and Anaplasma phagocytophilum: Rickettsiales pathogens of veterinary and public health significance’, Parasitology Research. Springer Verlag, pp. 3941–3957. Available at: https://doi.org/10.1007/s00436-015-4698-2. Baker, A. et al. (2020) ‘Increasing incidence of anaplasmosis in the United States, 2012 through 2016’, Vector-Borne and Zoonotic Diseases, 20(11), pp. 855–859. Available at: https://doi.org/10.1089/vbz.2019.2598. Bakken, J.S. and Dumler, J.S. (2015) ‘Human granulocytic anaplasmosis’, Infect Dis Clin of North Am., 29(1), pp. 341–355. Available at: https://doi.org/10.1016/j.idc.2015.02.007. Baldridge, G.D. et al. (2009) ‘Transovarial transmission of francisella-Like endosymbionts and Anaplasma phagocytophilum variants in Dermacentor albipictus (Acari: Ixodidae)’, Journal of Medical Entomology, 46(3), pp. 625–632. Available at: https://doi.org/10.1038/jid.2014.371. Barandika, J.F. et al. (2007) ‘Tick-borne zoonotic bacteria in wild and domestic small mammals in Northern Spain’, Applied and Environmental Microbiology, 73(19), pp. 6166–6171. Available at: https://doi.org/10.1128/AEM.00590-07. Barbet, A.F. et al. (2013) ‘An emerging tick-borne disease of humans is caused by a subset of strains with conserved genome structure’, Pathogens, 2(3), pp. 544–555. Available at: https://doi.org/10.3390/pathogens2030544. Beall, M.J. et al. (2008) ‘Serological and molecular prevalence of Borrelia burgdorferi, Anaplasma phagocytophilum, and Ehrlichia species in dogs from Minnesota’, Vector-Borne and Zoonotic Diseases, 8(4), pp. 455–464. Available at: https://doi.org/10.1089/vbz.2007.0236. Beninati, T. et al. (2006) ‘Anaplasmataceae in wild rodents and roe deer from Trento Province (northern Italy).’, Eur J Clin Microbiol Infect Dis, 25(10), pp. 677–678. Available at: https://doi.org/10.1007/s10096-006-0196-x. Beugnet, F. and Marié, J. Lou (2009) ‘Emerging arthropod-borne diseases of companion animals in Europe’, Veterinary Parasitology, 163(4), pp. 298–305. Available at: https://doi.org/10.1016/j.vetpar.2009.03.028. Biggs, H.M. et al. (2016) ‘Diagnosis and management of tickborne Rickettsial diseases: Rocky 29 Mountain spotted fever and other spotted fever group Rickettsioses, Ehrlichioses, and anaplasmosis -United States a practical guide for health care and public health professionals’. Centers for Disease Control and Prevention. Available at: http://www.cdc.gov/mmwr/cme/conted.html. Bjöersdorff, A. et al. (2001) ‘Ehrlichia-infected ticks on migrating birds’, Emerging Infectious Diseases, 7(5), pp. 877–879. Available at: https://doi.org/10.3201/eid0705.017517. Borjesson, D.L. et al. (2005) ‘Insights into pathogen immune evasion mechanisms: Anaplasma phagocytophilum fails to induce an apoptosis differentiation program in human neutrophils’, The Journal of Immunology, 174(10), pp. 6364–6372. Available at: https://doi.org/10.4049/jimmunol.174.10.6364. Bowman, D. et al. (2009) ‘Prevalence and geographic distribution of Dirofilaria immitis, Borrelia burgdorferi, Ehrlichia canis, and Anaplasma phagocytophilum in dogs in the United States: Results of a national clinic-based serologic survey’, Veterinary Parasitology, 160(1–2), pp. 138–148. Available at: https://doi.org/10.1016/j.vetpar.2008.10.093. Bown, K.J. et al. (2003) ‘Seasonal Dynamics of Anaplasma phagocytophila in a Rodent-Tick (Ixodes trianguliceps) System, United Kingdom’, Emerging Infectious Diseases •, 9(1), pp. 63– 70. Available at: https://doi.org/10.3201/eid0901.020169. Bown, K.J. et al. (2009) ‘Delineating Anaplasma phagocytophilum ecotypes in coexisting, discrete enzootic cycles’, Emerging Infectious Diseases, 15(12), pp. 1948–1954. Available at: https://doi.org/10.3201/eid1512.090178. Bray, D.P. et al. (2007) ‘Haemoparasites of common shrews (Sorex araneus) in Northwest England’, Parasitology, 134(6), pp. 819–826. Available at: https://doi.org/10.1017/S0031182007002302. Carrade, D.D. et al. (2009) ‘Canine granulocytic anaplasmosis: A review’, Journal of Veterinary Internal Medicine, pp. 1129–1141. Available at: https://doi.org/10.1111/j.1939- 1676.2009.0384.x. Carvalho, I.L. De et al. (2008) ‘Detection of Borrelia lusitaniae, Rickettsia sp. IRS3, Rickettsia monacensis, and Anaplasma phagocytophilum in Ixodes ricinus collected in Madeira Island, Portugal’, Vector-Borne and Zoonotic Diseases, 8(4), pp. 575–579. Available at: https://doi.org/10.1089/vbz.2007.0245 Caturegli, P. et al. (2000) ‘ankA: an Ehrlichia phagocytophila Group Gene Encoding a Cytoplasmic Protein Antigen with Ankyrin Repeats’, Infection and Immunity, 68(9), pp. 5277– 5283. Available at: https://doi.org/10.1128/IAI.68.9.5277-5283.2000. Chastagner, A. et al. (2014) ‘Multilocus sequence analysis of Anaplasma phagocytophilum reveals three distinct lineages with different host ranges in clinically ill French cattle’, Veterinary Research, 45(1), pp. 1–12. Available at: https://doi.org/10.1186/s13567-014-0114-7. 30 Chen, S.M. et al. (1994) ‘Identification of a granulocytotropic Ehrlichia species as the etiologic agent of human disease.’, Journal of Clinical Microbiology, 32(3), pp. 589–595. Available at: https://doi.org/10.1128/jcm.32.3.589-595.1994. Dahlgren, F.S. et al. (2015) ‘Human granulocytic anaplasmosis in the United States from 2008 to 2012: a summary of national surveillance data.’, The American journal of tropical medicine and hygiene, 93(1), pp. 66–72. Available at: https://doi.org/10.4269/ajtmh.15-0122. Daniels, T.J. et al. (2002) ‘Avian reservoirs of the agent of human granulocytic ehrlichiosis?’, Emerging Infectious Diseases, 8(12), pp. 1524–1525. Available at: https://doi.org/10.3201/eid0812.010527. Dingler, R.J. et al. (2014) ‘Surveillance for Ixodes pacificus and the tick-borne pathogens Anaplasma phagocytophilum and Borrelia burgdorferi in birds from California’s Inner Coast Range’, Ticks and Tick-borne Diseases, 5(4), pp. 436–445. Available at: https://doi.org/10.1016/j.ttbdis.2014.02.002. Drazenovich, N., Foley, J. and Brown, R.N. (2006) ‘Use of Real-time quantitative PCR targeting the msp2 protein gene to identify cryptic Anaplasma phagocytophilum infections in wildlife and domestic animals’, Vector Borne and Zoonotic Disease, 6(1), pp. 83–90. Available at: https://doi.org/10.1089/ast.2007.0153. Dugat, T. et al. (2015) ‘Opening the black box of Anaplasma phagocytophilum diversity: current situation and future perspectives.’, Frontiers in cellular and infection microbiology, 5, pp. 1–18. Available at: https://doi.org/10.3389/fcimb.2015.00061. Dumas, A. et al. (2022) ‘Transmission patterns of tick-borne pathogens among birds and rodents in a forested park in southeastern Canada’, PLoS ONE, 17(4). Available at: https://doi.org/10.1371/journal.pone.0266527. Dumic, I. et al. (2022) ‘Human Granulocytic Anaplasmosis—A Systematic Review of Published Cases’, Microorganisms. MDPI, pp. 1–15. Available at: https://doi.org/10.3390/microorganisms10071433. Dumitrache, M.O. et al. (2013) ‘Northern, white-breasted hedgehogs Erinaceus roumanicus as hosts for ticks infected with Borrelia burgdorferi sensu lato and Anaplasma phagocytophilum in Romania’, Ticks and Tick-borne Diseases, 4(3), pp. 214–217. Available at: https://doi.org/10.1016/j.ttbdis.2012.11.010. Dumler, J.S. (1997) ‘Is human granulocytic ehrlichiosis a new Lyme disease? Review and comparison of clinical, laboratory, epidemiological, and some biological features’, Clinical Infectious Diseases, 25(1 SUPPL.). Available at: https://doi.org/10.1086/516164. Dumler, J.S. et al. (2000) ‘Serum cytokine responses during acute human granulocytic ehrlichiosis’, Clinical and Diagnostic Laboratory Immunology, 7(1), pp. 6–8. Available at: 31 https://doi.org/10.1128/cdli.7.1.6-8.2000. Dumler, J.S. et al. (2001) ‘Reorganization of gene in families Rickettsiaceae and Anaplasmataceae in the order Rickettsiales: unification of some species of Ehrlichia with Anaplasma, Cowdria with Ehrlichia with Neorickettsia, description of six new species combinations and designation of Ehrlichia equi and “HGE agent” as subjective synonyms of Ehrlichia phagocytophila’, International Journal of systematic and evolutionary microbiology, 51(2001), pp. 2145–2165. Available at: https://doi.org/10.1099/00207713-51-6-2145. Dumler, J.S. et al. (2005) ‘Human granulocytic anaplasmosis and Anaplasma phagocytophilum’, Emerging Infectious Diseases, 11(12), pp. 1828–1834. Available at: https://doi.org/10.3201/eid1112.050898. Dunning Hotopp, J.C. et al. (2006) ‘Comparative genomics of emerging human ehrlichiosis agents’, PLoS Genetics, 2(2), pp. 208–223. Available at: https://doi.org/10.1371/journal.pgen.0020021. Elias, S.P. et al. (2020) ‘Surge in anaplasmosis cases in Maine, USA, 2013-2017’, Emerging Infectious Diseases, 26(2), pp. 327–331. Available at: https://doi.org/10.3201/eid2602.190529. Falco, Richard C., Durland, F. (1988) ‘Prevalence of Ixodes dammini near the homes of Lyme disease patients in Westchester County, New York’, American Journal of Epidemiology, 127(4), pp. 826–830. Available at: http://dx.doi.org/10.1093/oxfordjournals.aje.a114865. Földvári, G. et al. (2014) ‘Candidatus Neoehrlichia mikurensis and Anaplasma phagocytophilum in urban hedgehogs’, Emerging Infectious Diseases, 20(3), pp. 496–497. Available at: https://doi.org/10.3201/eid2003.130935. Foley, J. et al. (2008) ‘Possible differential host tropism in Anaplasma phagocytophilum strains in the western United States’, Annals of the New York Academy of Sciences, 1149, pp. 94–97. Available at: https://doi.org/10.1196/annals.1428.066. Foley, J. et al. (2016) ‘A putative marker for human pathogenic strains of Anaplasma phagocytophilum correlates with geography and host, but not human tropism’, Ticks and Tick- borne Diseases, 7(2), pp. 390–393. Available at: https://doi.org/10.1016/j.ttbdis.2015.12.015. Foley, J.E. et al. (2004) ‘Ecology of Anaplasma phagocytophilum and Borrelia burgdorferi in the western United States.’, Journal of Vector Ecology, 29(1), pp. 41–50. Available at: http://www.ncbi.nlm.nih.gov/pubmed/15266739. Foley, J.E. et al. (2008) ‘Anaplasma phagocytophilum infection in small mammal hosts of Ixodes ticks, Western United States’, Emerging Infectious Diseases, 14(7), pp. 1147–1150. Available at: https://doi.org/10.3201/eid1407.071599. Foley, J.E., Clueit, S.B. and Brown, R.N. (2008) ‘Differential exposure to Anaplasma phagocytophilum in rodent species in Northern California’, Vector-Borne and Zoonotic Diseases, 32 8(1), pp. 49–55. Available at: https://doi.org/10.1089/vbz.2007.0175. Foley, J.E. and Nieto, N.C. (2011) ‘The ecology of tick-transmitted infections in the redwood chipmunk (Tamias ochrogenys)’, Ticks and Tick-borne Diseases, 2(2), pp. 88–93. Available at: https://doi.org/10.1016/j.ttbdis.2010.11.003. Franke, J. et al. (2010) ‘Established and emerging pathogens in Ixodes ricinus ticks collected from birds on a conservation island in the Baltic Sea’, Medical and Veterinary Entomology, 24(4), pp. 425–432. Available at: https://doi.org/10.1111/j.1365-2915.2010.00905.x. Fuente, J. de la et al. (2005) ‘Potential Vertebrate Reservoir Hosts and Invertebrate Vectors of Anaplasma marginale and Anaplasma phagocytophilum in Central Spain’, Vector Borne and Zoonotic Disease, 5(4), pp. 390–401. Furman, D.P. and Loomis, E.C. (1984) ‘The ticks of California (Acari: Ixodida)’, in Bulletin of the California Insect Survey. Berkeley: University of California Press. Gabriel, M.W. et al. (2009) ‘Ecology of Anaplasma phagocytophilum infection in gray foxes (Urocyon cinereoargenteus) in northwestern California.’, Journal of wildlife diseases, 45(2), pp. 344–354. Available at: https://doi.org/10.7589/0090-3558-45.2.344. Ge, Y. et al. (2005) ‘Anaplasma phagocytophilum inhibits human neutrophil apoptosis via upregulation of bfl-1, maintenance of mitochondrial membrane potential and prevention of caspase 3 activation’, Cellular Microbiology, 7(1), pp. 29–38. Available at: https://doi.org/10.1111/j.1462-5822.2004.00427.x. Ge, Y. and Rikihisa, Y. (2007) ‘Identification of novel surface proteins of Anaplasma phagocytophilum by affinity purification and proteomics’, Journal of Bacteriology, 189(21), pp. 7819–7828. Available at: https://doi.org/10.1128/JB.00866-07. Grzeszczuk, A. et al. (2006) ‘The Root-vole Microtus oeconomus (Pallas, 1776): a new potential reservoir of Anaplasma phagocytophilum’, Vector Borne and Zoonotic Disease, 6(3), pp. 240– 243. Available at: https://doi.org/10.1089/vbz.2006.6.240. Hamer, S.A. et al. (2012) ‘Wild birds and urban ecology of ticks and tick-borne pathogens, Chicago, Illinois, USA, 2005-2010’, Emerging Infectious Diseases, 18(10), pp. 1589–1595. Available at: https://doi.org/10.3201/eid1810.120511. Khatat, S.E.H. et al. (2021) ‘Epidemiological and Clinicopathological Features of Anaplasma phagocytophilum Infection in Dogs: A Systematic Review’, Frontiers in Veterinary Science. Frontiers Media S.A. Available at: https://doi.org/10.3389/fvets.2021.686644. Herron, M.J. et al. (2000) ‘Intracellular parasitism by the human granulocytic ehrlichiosis bacterium through the P-selectin ligand, PSGL-1’, Science, 288(5471), pp. 1653–1656. Available at: https://doi.org/10.1126/science.288.5471.1653. 33 Hersh, M.H. et al. (2014) ‘When is a parasite not a parasite? Effects of larval tick burdens on white-footed mouse survival’, Ecology, 95(5), pp. 1360–1369. Available at: https://doi.org/10.1890/12-2156.1. Hildebrandt, A. et al. (2010) ‘The potential role of migratory birds in transmission cycles of Babesia spp., Anaplasma phagocytophilum, and Rickettsia spp.’, Ticks and Tick-borne Diseases, 1(2), pp. 105–107. Available at: https://doi.org/10.1016/j.ttbdis.2009.12.003. Hodzic, E. et al. (1998) ‘Acquisition and transmission of the agent of human granulocytic ehrlichiosis by Ixodes scapularis’, Journal of Clinical Microbiology, 36(12), pp. 3574–3578. Available at: https://doi.org/10.1128/jcm.36.12.3574-3578.1998. Hornok, S. et al. (2014) ‘Birds as potential reservoirs of tick-borne pathogens: First evidence of bacteraemia with Rickettsia helvetica’, Parasites and Vectors, 7(1), pp. 1–7. Available at: https://doi.org/10.1186/1756-3305-7-128. Huhn, C. et al. (2014) ‘Analysis of the population structure of Anaplasma phagocytophilum using multilocus sequence typing’, PLoS ONE. Edited by J.S. Dumler, 9(4), pp. 204–212. Available at: https://doi.org/10.1371/journal.pone.0093725. Hulínská, D. et al. (2004) ‘Detection of Anaplasma phagocytophilum in animals by real-time polymerase chain reaction’, Apmis, 112(4–5), pp. 239–247. Available at: https://doi.org/10.1111/j.1600-0463.2004.apm11204-0503.x. Ijdo, J.W., Carlson, A.C. and Kennedy, E.L. (2007) ‘Anaplasma phagocytophilum AnkA is tyrosine-phosphorylated at EPIYA motifs and recruits SHP-1 during early infection’, Cellular Microbiology, 9(5), pp. 1284–1296. Available at: https://doi.org/10.1111/j.1462- 5822.2006.00871.x. Jahfari, S. et al. (2014) ‘Circulation of four Anaplasma phagocytophilum ecotypes in Europe’, Parasites and Vectors, 7(1), pp. 1–11. Available at: https://doi.org/10.1186/1756-3305-7-365. Jin, H. et al. (2012) ‘Epidemiology and control of human granulocytic anaplasmosis: a systematic review’, Vector-Borne and Zoonotic Diseases, 12(4), pp. 269–274. Available at: https://doi.org/10.1089/vbz.2011.0753. Johnson, R.C. et al. (2011) ‘Agents of human anaplasmosis and Lyme Disease at Camp Ripley, Minnesota’, Vector-Borne and Zoonotic Diseases, 11(12), pp. 1529–1534. Available at: https://doi.org/10.1089/vbz.2011.0633. Johnston, E. et al. (2013) ‘Anaplasma phagocytophilum Infection in American Robins and Gray Catbirds: An Assessment of Reservoir Competence and Disease in Captive Wildlife’, Journal of Medical Entomology, 50(1), pp. 163–170. Available at: https://doi.org/10.1603/ME12141. Kauffmann, M. et al. (2017) ‘Anaplasma phagocytophilum and Babesia spp. in roe deer (Capreolus capreolus), fallow deer (Dama dama) and mouflon (Ovis musimon) in Germany’, 34 Molecular and Cellular Probes, 31, pp. 46–54. Available at: https://doi.org/10.1016/j.mcp.2016.08.008. Keesing, F. et al. (2009) ‘Hosts as ecological traps for the vector of Lyme disease’, Proceedings of the Royal Society B: Biological Sciences, 276(1675), pp. 3911–3919. Available at: https://doi.org/10.1098/rspb.2009.1159. Keesing, F. et al. (2012) ‘Reservoir Competence of Vertebrate Hosts for Anaplasma phagocytophilum’, Emerging Infectious Diseases, 18(12), pp. 10–13. Available at: https://doi.org/10.3201/eid1812.120919. Keesing, F. et al. (2014) ‘Prevalence of human-Active and variant 1 strains of the tick-borne pathogen Anaplasma phagocytophilum in hosts and forests of Eastern North America’, American Journal of Tropical Medicine and Hygiene, 91(2), pp. 302–309. Available at: https://doi.org/10.4269/ajtmh.13-0525. Ladbury, G.A.F. et al. (2008) ‘Dynamic transmission of numerous Anaplasma phagocytophilum genotypes among lambs in an infected sheep flock in an area of anaplasmosis endemicity’, Journal of Clinical Microbiology, 46(5), pp. 1686–1691. Available at: https://doi.org/10.1128/JCM.02068-07. Lagler, H. et al. (2017) ‘Direct detection of Anaplasma phagocytophilum by polymerase chain reaction followed by electrospray ionization mass spectrometry from human blood’, International Journal of Infectious Diseases, 60, pp. 61–63. Available at: https://doi.org/10.1016/j.ijid.2017.05.006. Levin, M.L. et al. (2002) ‘Comparison of reservoir competence of medium sized mammals and Peromyscus leucopus for Anaplasma phagocytophilum in Connecticut’, Vector Borne and Zoonotic Disease, 2(3). Available at: https://doi.org/10.1089/ast.2007.0153. Levin, M.L. et al. (2021) ‘Incompetence of the Asian long horned tick (Acari: Ixodidae) in transmitting the agent of human granulocytic anaplasmosis in the United States’, Journal of Medical Entomology, 58(3), pp. 1419–1423. Available at: https://doi.org/10.1093/jme/tjab015. Levin, M.L. and Fish, D. (2000a) ‘Acquisition of coinfection and simultaneous transmission of Borrelia burgdorferi and Ehrlichia phagocytophila by Ixodes scapularis ticks’, Infection and Immunity, 68(4), pp. 2183–2186. Available at: https://doi.org/10.1128/IAI.68.4.2183-2186.2000. Levin, M.L. and Fish, D. (2000b) ‘Immunity reduces reservoir host competence of Peromyscus leucopus for Ehrlichia phagocytophila’, Infection and Immunity, 68(3), pp. 1514–1518. Available at: https://doi.org/10.1128/IAI.68.3.1514-1518.2000. Levin, M.L. and Ross, D.E. (2004) ‘Acquisition of different isolates of Anaplasma phagocytophilum by Ixodes scapularis from a model animal.’, Vector-borne and zoonotic diseases, 4(1), pp. 53–59. Available at: https://doi.org/10.1089/153036604773082997. 35 Levin, M.L., Troughton, D.R. and Loftis, A.D. (2021) ‘Duration of tick attachment necessary for transmission of Anaplasma phagocytophilum by Ixodes scapularis (Acari: Ixodidae) nymphs’, Ticks and Tick-borne Diseases, 12(6). Available at: https://doi.org/10.1016/j.ttbdis.2021.101819. Lin, M. and Rikihisa, Y. (2003a) ‘Ehrlichia chaffeensis and Anaplasma phagocytophilum lack genes for lipid A biosynthesis and incorporate cholesterol for their survival’, Infection and Immunity, 71(9), pp. 5324–5331. Available at: https://doi.org/10.1128/IAI.71.9.5324-5331.2003. Lin, M. and Rikihisa, Y. (2003b) ‘Obligatory intracellular parasitism by Ehrlichia chaffeensis and Anaplasma phagocytophilum involves caveolae and glycosylphosphatidylinositol-anchored proteins’, Cellular Microbiology, 5(11), pp. 809–820. Available at: https://doi.org/10.1046/j.1462-5822.2003.00322.x. Little, S.E. (2010) ‘Ehrlichiosis and Anaplasmosis in Dogs and Cats’, Veterinary Clinics of North America - Small Animal Practice, 40(6), pp. 1121–1140. Available at: https://doi.org/10.1016/j.cvsm.2010.07.004. Liu, L. et al. (2011) ‘Ixodes scapularis salivary gland protein P11 facilitates migration of Anaplasma phagocytophilum from the tick gut to salivary glands’, EMBO Rep., 12(11), pp. 1196–1203. Available at: https://doi.org/10.1038/embor.2011.177. Liz, J.S. et al. (2001) ‘PCR detection of granulocytic ehrlichiae in Ixodes ricinus ticks and wild small mammals in Western Switzerland’, Journal of Clinical Microbiology, 39(2), p. 828. Loewenich, F.D. Von et al. (2003) ‘High Diversity of ankA Sequences of Anaplasma phagocytophilum among Ixodes ricinus Ticks in Germany’, 41(11), pp. 5033–5040. Available at: https://doi.org/10.1128/JCM.41.11.5033. Magnarelli, L.A. et al. (1997) ‘Antibodies to multiple tick-borne pathogens of babesiosis, ehrlichiosis, and Lyme borreliosis in white-footed mice’, Journal of Wildlife Diseases, 33(3), pp. 466–473. Available at: https://doi.org/10.7589/0090-3558-33.3.466. Majazki, J. et al. (2013) ‘Anaplasma phagocytophilum strains from voles and shrews exhibit specific ankA gene sequences.’, BMC veterinary research, 9, p. 235. Available at: https://doi.org/10.1186/1746-6148-9-235. Mărcuţan, I. et al. (2014) ‘Prevalence of Anaplasma phagocytophilum in ticks collected from migratory birds in Danube Delta, Romania’, Parasites & Vectors, 7(Suppl 1), p. P16. Available at: https://doi.org/10.1186/1756-3305-7-s1-p16. Marumoto, K. et al. (2007) ‘Detection of Anaplasma phagocytophilum and Ehrlichia sp. HF strains in Ixodes ricinus ticks in Brittany, France’, Clinical Microbiology and Infection, 13(3), pp. 338–341. Available at: https://doi.org/10.1111/j.1469-0691.2006.01630.x. Massung, R.F. et al. (1998) ‘Nested PCR Assay for Detection of Granulocytic Ehrlichiae’, Journal of Clinical Microbiology, 36(4), pp. 1090–1095. Available at: 36 https://doi.org/10.1128/jcm.36.4.1090-1095.1998. Massung, R.F. et al. (2002) ‘Genetic variants of Ehrlichia phagocytophila, Rhode Island and Connecticut’, Emerging Infectious Diseases, 8(5), pp. 467–472. Available at: https://doi.org/10.3201/eid0805.010251. Massung, R.F. et al. (2003) ‘Inability of a variant strain of Anaplasma phagocytophilum to infect mice.’, The Journal of infectious diseases, 188(11), pp. 1757–63. Available at: https://doi.org/10.1086/379725. Massung, R.F. et al. (2005) ‘Anaplasma phagocytophilum in white-tailed deer’, Emerging Infectious Diseases, 11(10), pp. 1604–1606. Available at: https://doi.org/10.3201/eid1110.041329. McQuiston, J.H. et al. (1999) ‘The human Ehrlichioses in the United States’, Emerging Infectious Diseases, 5(5), pp. 635–642. Available at: https://doi.org/10.3201/eid0505.990504. Michalski, M. et al. (2006) ‘Anaplasma phagocytophilum in central and western Wisconsin: A molecular survey’, Parasitology Research, 99(6), pp. 694–699. Available at: https://doi.org/10.1007/s00436-006-0217-9. Morissette, E. et al. (2009) ‘Diversity of Anaplasma phagocytophilum strains, USA’, Emerging Infectious Diseases, 15(6), pp. 928–931. Available at: https://doi.org/10.3201/eid1506.081610. Mott, J., Barnewall, R.E. and Rikihisa, Y. (1999) ‘Human granulocytic Ehrlichiosis agent and Ehrlichia chaffeensis reside in different cytoplasmic compartments in HL-60 cells’, Infection and Immunity, 67(3), pp. 1368–1378. Available at: https://doi.org/10.1128/iai.67.3.1368-1378.1999. Mukhacheva, T.A. et al. (2020) ‘Phylogeographical diversity of Anaplasma phagocytophilum in the Asian part of Russia based on multilocus sequence typing and analysis of the ankA gene’, Infection, Genetics and Evolution, 80(February), p. 104234. Available at: https://doi.org/10.1016/j.meegid.2020.104234. Murphy, D.S. et al. (2017) ‘Prevalence and distribution of human and tick infections with the Ehrlichia muris -Like agent and Anaplasma phagocytophilum in Wisconsin, 2009–2015’, Vector- Borne and Zoonotic Diseases, 17(4), pp. 229–236. Available at: https://doi.org/10.1089/vbz.2016.2055. Musser, G.G. (1969) ‘King, John A., (ed.). Biology of Peromyscus (Rodentia). Spec. Publ. 2, Amer. Soc. Mammalogists, xiii + 594 pp, frontispiece, 63 figs., December 20, 1968. Price, $15.00’, Journal of Mammalogy, 50(3), p. 655. Available at: https://doi.org/https://doi.org/10.2307/1378817. Neelakanta, G. et al. (2010) ‘Anaplasma phagocytophilum induces Ixodes scapularis ticks to express an antifreeze glycoprotein gene that enhances their survival in the cold’, Journal of Clinical Investigation, 120(9), pp. 3179–3190. Available at: https://doi.org/10.1172/JCI42868. 37 Nicholls, T.H. and Callister, S.M. (1996) ‘Lyme disease spirochetes in ticks collected from birds in Midwestern United States’, Journal of Medical Entomology, 33(3), pp. 379–384. Available at: https://doi.org/10.1093/jmedent/33.3.379. Nicholson, W.L. et al. (1999) ‘Dusky-footed wood rats (Neotoma fuscipes) as reservoirs of granulocytic ehrlichiae (Rickettsiales: Ehrlichieae) in northern California’, Journal of Clinical Microbiology, 37(10), pp. 3323–3327. Available at: https://doi.org/10.1128/jcm.37.10.3323- 3327.1999. Nieto, N.C. et al. (2009) ‘Reptile infection with Anaplasma phagocytophilum, the causative agent of granulocytic anaplasmosis.’, The Journal of parasitology, 95(5), pp. 1165–70. Available at: https://doi.org/10.1645/GE-1983.1. Nieto, N.C. and Foley, J.E. (2008) ‘Evaluation of Squirrels (Rodentia: Sciuridae) as Ecologically Significant Hosts for Anaplasma phagocytophilum in California’, Journal of Medical Entomology, 45(4), pp. 763–769. Available at: https://doi.org/10.1603/0022- 2585(2008)45[763:EOSRSA]2.0.CO;2. Niu, H., Yamaguchi, M. and Rikihisa, Y. (2008) ‘Subversion of cellular autophagy by Anaplasma phagocytophilum’, Cellular Microbiology, 10(3), pp. 593–605. Available at: https://doi.org/10.1111/j.1462-5822.2007.01068.x. Ogden, N.H. et al. (1998) ‘Granulocytic Ehrlichia infection in Ixodid ticks and mammals in woodlands and uplands of the U.K.’, Medical and Veterinary Entomology, 12(4), pp. 423–429. Available at: https://doi.org/10.1046/j.1365-2915.1998.00133.x. Ogden, N.H. et al. (2002) ‘Transmission of Anaplasma phagocytophilum to Ixodes ricinus Ticks from Sheep in the Acute and Post-Acute Phases of Infection’, Infection and Immunity, 71(4), pp. 2071–2078. Available at: https://doi.org/10.1128/IAI.71.4.2071-2078.2003. Ogden, N.H. et al. (2008) ‘Role of migratory birds in introduction and range expansion of Ixodes scapularis ticks and of Borrelia burgdorferi and Anaplasma phagocytophilum in Canada’, Applied and Environmental Microbiology, 74(6), pp. 1780–1790. Available at: https://doi.org/10.1128/AEM.01982-07. Oporto, B. et al. (2003) ‘A survey on Anaplasma phagocytophila in wild small mammals and roe deer (Capreolus capreolus) in Northern Spain’, Annals of the New York Academy of Sciences, 990, pp. 98–102. Available at: https://doi.org/10.1111/j.1749-6632.2003.tb07344.x. Overzier, E. et al. (2013) ‘Detection of tick-borne pathogens in roe deer (Capreolus capreolus), in questing ticks (Ixodes ricinus), and in ticks infesting roe deer in southern Germany’, Ticks and Tick-borne Diseases, 4(4), pp. 320–328. Available at: https://doi.org/10.1016/j.ttbdis.2013.01.004. Palomar, A.M. et al. (2012) ‘Role of birds in dispersal of etiologic agents of tick-borne zoonoses, Spain, 2009’, Emerging Infectious Diseases, 18(7), pp. 1188–1191. Available at: 38 https://doi.org/10.3201/eid1807.111777. Park, J. et al. (2004) ‘Anaplasma phagocytophilum AnkA binds to granulocyte DNA and nuclear proteins’, Cellular Microbiology, 6(8), pp. 743–751. Available at: https://doi.org/10.1111/j.1462- 5822.2004.00400.x. Paulauskas, A., Radzijevskaja, J. and Rosef, O. (2009) ‘Anaplasma in ticks feeding on migrating birds and questing ticks in Lithuania and Norway’, Clinical Microbiology and Infection, 15(Supplement 2), pp. 34–36. Available at: https://doi.org/https://doi.org/10.1111/j.1469- 0691.2008.02164.x. Price, K.J. et al. (2022) ‘First detection of human pathogenic variant of Anaplasma phagocytophilum in field-collected Haemaphysalis longicornis, Pennsylvania, USA’, Zoonoses and Public Health, 69(2), pp. 143–148. Available at: https://doi.org/10.1111/zph.12901. Qurollo, B.A. et al. (2014) ‘A serological survey of tick-borne pathogens in dogs in North America and the Caribbean as assessed by Anaplasma phagocytophilum, A. platys, Ehrlichia canis, E. chaffeensis, E. ewingii, and Borrelia burgdorferi species-specific peptides’, Infection Ecology & Epidemiology, 4(1), p. 24699. Available at: https://doi.org/10.3402/iee.v4.24699. Rar, V. and Golovljova, I. (2011) ‘Anaplasma, Ehrlichia, and “Candidatus Neoehrlichia” bacteria: Pathogenicity, biodiversity, and molecular genetic characteristics, a review’, Infection, Genetics and Evolution, 11(8), pp. 1842–1861. Available at: https://doi.org/10.1016/j.meegid.2011.09.019. Rar, V., Tkachev, S. and Tikunova, N. (2021) ‘Genetic diversity of Anaplasma bacteria: Twenty years later’, Infection, Genetics and Evolution, 91. Available at: https://doi.org/10.1016/j.meegid.2021.104833. Remesar, S. et al. (2020) ‘Prevalence and molecular characterization of Anaplasma phagocytophilum in roe deer (Capreolus capreolus) from Spain’, Ticks and Tick-borne Diseases, 11(2), p. 101351. Available at: https://doi.org/10.1016/j.ttbdis.2019.101351. Rikihisa, Y. (2010) ‘Anaplasma phagocytophilum and Ehrlichia chaffeensis: Subversive manipulators of host cells’, Nature Reviews Microbiology, 8(5), pp. 328–339. Available at: https://doi.org/10.1038/nrmicro2318. Rikihisa, Y. (2011) ‘Mechanisms of obligatory intracellular infection with Anaplasma phagocytophilum’, Clinical Microbiology Reviews, 24(3), pp. 469–489. Available at: https://doi.org/10.1128/CMR.00064-10. Scaife, H. et al. (2003) ‘Anaplasma phagocytophilum reduces neutrophil apoptosis in vivo’, Infection and Immunity, 71(4), pp. 1995–2001. Available at: https://doi.org/10.1128/IAI.71.4.1995-2001.2003. Scharf, W. et al. (2011) ‘Distinct host species correlate with Anaplasma phagocytophilum ankA 39 gene clusters’, Journal of Clinical Microbiology, 49(3), pp. 790–796. Available at: https://doi.org/10.1128/JCM.02051-10. Scharf, W.C. (2004) ‘Immature ticks on birds: Temporal abundance and reinfestation’, Northeastern Naturalist, 11(2), pp. 143–150. Available at: https://doi.org/10.1656/1092- 6194(2004)011[0143:ITOBTA]2.0.CO;2. Scott, J.D., Anderson, J.F. and Durden, L.A. (2012) ‘Widespread dispersal of Borrelia burgdorferi-infected ticks collected from songbirds across Canada’, Journal of Parasitology, 98(1), pp. 49–59. Available at: https://doi.org/10.1645/GE-2874.1. Severo, M.S. et al. (2012) ‘Anaplasma phagocytophilum: deceptively simple or simply deceptive?’, Future Microbiology, 7, pp. 719–731. Available at: https://doi.org/10.2217/fmb.12.45. Silaghi, C., Woll, D., et al. (2012) ‘Babesia spp. and Anaplasma phagocytophilum in questing ticks, ticks parasitizing rodents and the parasitized rodents--analyzing the host-pathogen-vector interface in a metropolitan area.’, Parasites & vectors, 5(1), p. 191. Available at: https://doi.org/10.1186/1756-3305-5-191. Silaghi, C., Skuballa, J., et al. (2012) ‘The European hedgehog (Erinaceus europaeus) - A suitable reservoir for variants of Anaplasma phagocytophilum?’, Ticks and Tick-borne Diseases, 3(1), pp. 49–54. Available at: https://doi.org/10.1016/j.ttbdis.2011.11.005. Silaghi, C. et al. (2020) ‘Anaplasma phagocytophilum and Babesia Species of Sympatric Roe Deer (Capreolus capreolus), Fallow Deer (Dama dama), Sika Deer (Cervus nippon) and Red Deer (Cervus elaphus) in Germany’, Pathogens, 9, pp. 1–11. Silaghi, C., Liebisch, G. and Pfister, K. (2011) ‘Genetic variants of Anaplasma phagocytophilum from 14 equine granulocytic anaplasmosis cases’, Parasites and Vectors, 4(1), p. 161. Available at: https://doi.org/10.1186/1756-3305-4-161. Skotarczak, B. et al. (2006) ‘PCR detection of granulocytic Anaplasma and Babesia in Ixodes ricinus ticks and birds in west-central Poland’, Annals of Agricultural and Environmental Medicine, 13(1), pp. 21–23. Available at: https://doi.org/10.30701/ijc.v39isuppl_b.854. Smetanová, K., Schwarzová, K. and Kocianová, E. (2006) ‘Detection of Anaplasma phagocytophilum, Coxiella burnetii, Rickettsia spp., and Borrelia burgdorferi s. l. in ticks, and wild-living animals in Western and Middle Slovakia’, Annals of the New York Academy of Sciences, 1078, pp. 312–315. Available at: https://doi.org/10.1196/annals.1374.058. Smith, R.P. et al. (1996) ‘Role of bird migration in the long-distance dispersal of Ixodes dammini, the vector of Lyme disease’, Journal of Infectious Diseases, 174(1), pp. 221–224. Available at: https://doi.org/10.1093/infdis/174.1.221. Stafford, K.C. et al. (1999) ‘Infection with agents of human granulocytic ehrlichiosis, Lyme 40 disease, and babesiosis in wild white-footed mice (Peromyscus leucopus) in Connecticut’, 37(9), pp. 2887–2892. Available at: https://doi.org/10.1128/jcm.37.9.2887-2892.1999. Stafford, K.C., Bladen, V.C. and Magnarelli, L.A. (1995) ‘Ticks (Acari: Ixodidae) infesting wild birds (Aves) and white-footed mice in Lyme, CT’, Journal of Medical Entomology, 32(4), pp. 453–466. Available at: https://doi.org/10.1093/jmedent/32.4.453. Stephenson, N. and Foley, J. (2016) ‘Parallelisms and Contrasts in the Diverse Ecologies of the Anaplasma phagocytophilum and Borrelia burgdorferi Complexes of Bacteria in the Far Western United States’, Veterinary Sciences, 3(4), p. 26. Available at: https://doi.org/10.3390/vetsci3040026. Stigum, V.M. et al. (2019) ‘Infection prevalence and ecotypes of Anaplasma phagocytophilum in moose Alces alces, red deer Cervus elaphus, roe deer Capreolus capreolus and Ixodes ricinus ticks from Norway’, Parasites and Vectors, 12(1), pp. 1–8. Available at: https://doi.org/10.1186/s13071-018-3256-z. Stuen, S. et al. (2002) ‘Identification of Anaplasma phagocytophila (formerly Ehrlichia phagocytophila) variants in blood from sheep in Norway’, Journal of Clinical Microbiology, 40(9), pp. 3192–3197. Available at: https://doi.org/10.1128/JCM.40.9.3192-3197.2002. Stuen, S. et al. (2003) ‘Differences in clinical manifestations and hematological and serological responses after experimental infection with genetic variants of Anaplasma phagocytophilum in sheep’, Clinical and Diagnostic Laboratory Immunology, 10(4), pp. 692–695. Available at: https://doi.org/10.1128/CDLI.10.4.692-695.2003. Stuen, S. (2007) ‘Anaplasma Phagocytophilum - The most widespread tick-borne infection in animals in Europe’, Veterinary Research Communications, 31(SUPPL. 1), pp. 79–84. Available at: https://doi.org/10.1007/s11259-007-0071-y. Stuen, S. et al. (2013) ‘Anaplasma phagocytophilum variants in sympatric red deer (Cervus elaphus) and sheep in southern Norway’, Ticks and Tick-borne Diseases, 4(3), pp. 197–201. Available at: https://doi.org/10.1016/j.ttbdis.2012.11.014. Stuen, S., Granquist, E.G. and Silaghi, C. (2013) ‘Anaplasma phagocytophilum--a widespread multi-host pathogen with highly adaptive strategies.’, Frontiers in cellular and infection microbiology, 3(July), p. 31. Available at: https://doi.org/10.3389/fcimb.2013.00031. Sukumaran, B. et al. (2006) ‘An Ixodes scapularis protein required for survival of Anaplasma phagocytophilum in tick salivary glands.’, The Journal of experimental medicine, 203(6), pp. 1507–17. Available at: https://doi.org/10.1084/jem.20060208. Tak, T. et al. (2013) ‘Whatˈs your age again? Determination of human neutrophil half-lives revisited’, Journal of Leukocyte Biology, 94(4), pp. 595–601. Available at: https://doi.org/10.1189/jlb.1112571. 41 Telford, S.R. et al. (1996) ‘Perpetuation of the agent of human granulocytic ehrlichiosis in a deer tick-rodent cycle’, Proceedings of the National Academy of Sciences of the United States of America, 93(12), pp. 6209–6214. Available at: https://doi.org/10.1073/pnas.93.12.6209. Troese, M.J. and Carlyon, J.A. (2009) ‘Anaplasma phagocytophilum dense-cored organisms mediate cellular adherence through recognition of human P-selectin glycoprotein ligand 1’, Infection and Immunity, 77(9), pp. 4018–4027. Available at: https://doi.org/10.1128/IAI.00527- 09. Trost, C.N. et al. (2018) ‘Three genetically distinct clades of Anaplasma phagocytophilum in Ixodes scapularis’, Ticks and Tick-borne Diseases, 9(6), pp. 1518–1527. Available at: https://doi.org/10.1016/j.ttbdis.2018.07.002. Troughton, D.R. and Levin, M.L. (2007) ‘Life Cycles of Seven Ixodid Tick Species (Acari: Ixodidae) Under Standardized Laboratory Conditions’, J. Med. Entomol, 44(5), pp. 732–740. Available at: https://doi.org/https://doi.org/10.1093/jmedent/44.5.732. Tsiodras, S. et al. (2017) ‘Fatal human anaplasmosis associated with macrophage activation syndrome in Greece and the Public Health response’, Journal of Infection and Public Health, 10(6), pp. 819–823. Available at: https://doi.org/10.1016/j.jiph.2017.01.002. Veronesi, F. et al. (2011) ‘Prevalence of Anaplasma phagocytophilum in fallow deer (Dama dama) and feeding ticks from an Italy preserve’, Research in Veterinary Science, 90(1), pp. 40– 43. Available at: https://doi.org/10.1016/j.rvsc.2010.05.019. des Vignes, F. et al. (2001) ‘Effect of tick removal on transmission of Borrelia burgdorferi and Ehrlichia phagocytophila by Ixodes scapularis Nymphs’, Journal of Infectious Diseases, 183(5), pp. 773–778. Available at: https://doi.org/10.1086/318818. Walls, J.J. et al. (1997) ‘Natural infection of small mammal species in Minnesota with the agent of human granulocytic ehrlichiosis.’, Journal of clinical microbiology, 35(4), pp. 853–5. Available at: https://doi.org/10.1128/jcm.35.4.853-855.1997. Wang, X., Kikuchi, T. and Rikihisa, Y. (2006) ‘Two monoclonal antibodies with defined epitopes of P44 major surface proteins neutralize Anaplasma phagocytophilum by distinct mechanisms’, Infection and Immunity, 74(3), pp. 1873–1882. Available at: https://doi.org/10.1128/IAI.74.3.1873-1882.2006. Webster, P. et al. (1998) ‘The agent of human granulocytic ehrlichiosis resides in an endosomal compartment’, Journal of Clinical Investigation, 101(9), pp. 1932–1941. Available at: https://doi.org/10.1172/JCI1544. Woldehiwet, Z. (2010) ‘The natural history of Anaplasma phagocytophilum’, Veterinary Parasitology, 167(2–4), pp. 108–122. Available at: https://doi.org/10.1016/j.vetpar.2009.09.013. Yabsley, M.J. et al. (2008) ‘Experimental and field studies on the suitability of raccoons 42 (Procyon lotor) as hosts for tick-borne pathogens.’, Vector borne and zoonotic diseases, 8(4), pp. 491–503. Available at: https://doi.org/10.1089/vbz.2007.0240. Yoshiie, K. et al. (2000) ‘Intracellular infection by the human granulocytic ehrlichiosis agent inhibits human neutrophil apoptosis’, Infection and Immunity, 68(3), pp. 1125–1133. Available at: https://doi.org/10.1128/IAI.68.3.1125-1133.2000. Zeman, P. and Pecha, M. (2008) ‘Segregation of genetic variants of Anaplasma phagocytophilum circulating among wild ruminants within a Bohemian Forest (Czech Republic)’, International Journal of Medical Microbiology, 298(SUPPL. 1), pp. 203–210. Available at: https://doi.org/10.1016/j.ijmm.2008.03.003. 43 CHAPTER 3: THE ECOLOGY OF ANAPLASMA PHAGOCYTOPHILUM AT A SITE IN THE UPPER MIDWEST USA (FORT MCCOY, WISCONSIN) ABSTRACT Anaplasma phagocytophilum is an intracellular bacterium that causes canine and human granulocytic anaplasmosis, and most cases in the United States occur in two main foci: the Upper Midwest and the Northeast. In these regions, A. phagocytophilum is vectored by the blacklegged tick (= deer tick), Ixodes scapularis Say 1821. Although human granulocytic anaplasmosis is the second most common vector-borne disease in the US, few ecological studies have been conducted in the Upper Midwest. The main objective of this study was to characterize the ecology of A. phagocytophilum at a site in central Wisconsin where blacklegged ticks have been established for more than half a century and human anaplasmosis is endemic. Sampling was conducted from May – September in 2010 – 2012 every 2-3 weeks during which questing ticks and small and medium- sized animals were live-captured on 3 1-ha grids. A subset of questing ticks, attached ticks, host biopsies, and host blood were assayed for infection with A. phagocytophilum by a real time PCR followed by a confirmatory nested PCR. Over the three years, 1284 number ticks were collected; the average infection prevalence of questing nymphs and adults for Ap-ha was 15.3% and 20.4% respectively. All infected ticks were infected with the Ap-ha strain. The infection prevalence of A. phagocytophilum of small to medium sized mammals comprising 10 species (N = 1244 individuals) ranged from 5.7% to 89%. Peromyscus leucopus (white-footed mouse) was the most frequently captured species (N = 571, infection prevalence = 28.0%) and Tamias striatus (eastern chipmunk) had the highest infection prevalence (88.9%). Based on the fully engorged larvae collected from white-footed mice and eastern chipmunks, the realized reservoir competence of eastern chipmunks (88.9%) was significantly greater than white-footed mice (12.9%), but white- 44 footed mice fed a significantly greater number of larvae (71.7% of the total larvae collected from animals) compared to eastern chipmunks (1.2% of the total larvae collected from animals). Therefore, the relative contribution of white-footed mice to the enzootic cycle of A. phagocytophilum may be more significant at this field site. The phenology of infection prevalence of hosts and on-host larvae showed a similar pattern where infection prevalence increased from May to June and then decreased. This pattern is comparable with the phenology of questing infected I. scapularis nymphs. We also compared the infection prevalence between blood and ear biopsies of white-footed mice (N=367) that were obtained during the same capture event. There was no significant difference in infection prevalence in mice from which A. phagocytophilum was collected from blood only, biopsy only, and both types simultaneously for each pairwise analysis. Interestingly, the infection prevalence of on-host larvae was highest when collected from mice from which A. phagocytophilum was detected in both ear biopsies and blood samples at the same capture event. Future ecological studies should be conducted across the Midwest to assess similarities and differences in the enzootic maintenance cycle; furthermore, they should also explore the roles of birds and other hosts not captured (or not thoroughly studied here), making certain to differentiate among A. phagocytophilum strains. Future research may also look deeper into the infection, transmission, and immunity dynamics of highly infected hosts such as the eastern chipmunk to better understand the host-pathogen interactions. Finally, laboratory-based studies should also compare the temporal dynamics of detection of A. phagocytophilum from ear, blood, and transmission to feeding larvae in xenodiagnoses experiments to better understand the relevance of detection of A. phagocytophilum in blood or biopsy samples given the logistical advantages for collecting biopsies in field work. 45 Keywords: Anaplasma phagocytophilum, Upper Midwest, blacklegged tick, white-footed mouse, eastern chipmunk 46 INTRODUCTION Anaplasma phagocytophilum is an intracellular bacterium and is known to cause granulocytic anaplasmosis in humans, equine granulocytic anaplasmosis in horses, febrile fever in cats, and canine granulocytic anaplasmosis in dogs (Chen et al., 1994; Stuen, Granquist and Silaghi, 2013; Dugat et al., 2015). Human granulocytic anaplasmosis is an infectious zoonotic disease in North America, Europe, and Asia (Stuen, 2007; Stuen, Granquist and Silaghi, 2013; Silaghi et al., 2017). In the United States, after Lyme disease, anaplasmosis is the second leading vector-borne and tick-borne disease. Furthermore, like Lyme disease, the reported human case incidence for anaplasmosis is greatest in the Northeast and the Upper Midwest (Dahlgren et al., 2015). Anaplasma phagocytophilum is vectored by ticks that belong to the Ixodes ricinus complex, the same species complex that vector the agents of Lyme disease in North America and Eurasia. In the US, as with the Lyme disease agent Borrelia burgdorferi, A. phagocytophilum is vectored by Ixodes pacificus in the western and Ixodes scapularis in the eastern USA (Barlough et al., 1997; Pancholi et al., 1995; Richter Jr. et al., 1996; Telford et al., 1996). In the US, there are two major strains of A. phagocytophilum circulating within the same enzootic cycle, one strain is the human pathogenic strain Ap-ha, which causes disease in humans, and the other strain is the deer variant Ap-v1, which is not known to cause disease in humans. Although many studies have been conducted to elucidate the eco-epidemiology of Lyme disease, less effort has been focused on that of human anaplasmosis. Less research also has been conducted on the ecology of the agent of anaplasmosis in the Upper Midwest compared to in either the northeastern or western US. Limited data suggest that the ecology of the enzootic cycle maintaining A. phagocytophilum in the Upper Midwest US should be similar to that in the 47 Northeast, but that there may be differences based on regional differences in the ecology of I. scapularis (e.g., differences in phenology, Ogden et al. 2007; Gatewood et al. 2009; Ogden et al. 2018). Thus, using samples from a study originally designed to study the ecology of the Lyme disease pathogen, we had the opportunity to better characterize the ecology of Ap-ha in central Wisconsin, an area in the Upper Midwest highly endemic for I. scapularis-borne diseases. Our objectives were: 1) to estimate the prevalence of infection of questing I. scapularis (nymphs and adults) for Ap-ha; 2) to estimate the prevalence of infection of small and medium-sized mammals for Ap-ha; 3) to characterize the realized reservoir competence for Ap-ha of small and medium- sized mammals often parasitized by I. scapularis; and 4) to characterize the seasonal enzootic dynamics of infection between ticks and hosts. As part of this work, we also compare the use of blood versus ear biopsies for inferring the infection and transmission status of wildlife. MATERIALS AND METHODS Field site and study design We investigated the ecology of I. scapularis and A. phagocytophilum at Fort McCoy Military Installation, in Monroe County, western central Wisconsin (WI; 44.0391 N, −90.6766 W). Ixodes scapularis, B. burgdorferi and Lyme disease have been endemic at Fort McCoy for decades (Anderson, Duray and Magnarelli, 1987). Anaplasma phagocytophilum has been detected there previously (Steiner et al. 2008; Hamer et al. 2014). The field site, experimental design, and sampling protocols have been described previously in Ginsberg et al., 2021, Ogden et al., 2018, Rulison et al., 2013. Below we summarize the most pertinent aspects and refer the reader to prior studies for additional details. Sampling was carried out on three 1-ha grids that are dominated by oak (Quercus spp.), pines (Pinus spp.), red maples (Acer rubrum), and have a shrub layer consisting of mainly tree saplings (Rulison et al., 2013; Arsnoe et al., 2015). Each grid was 48 separated by ≥ 3 km. Grids were sampled either every two (2010) or three (2011-2012) weeks from May to September, with some additional sampling outside that period. Questing ticks Sampling for questing ticks was conducted on rain-free days. Questing ticks were collected by a combination of dragging and flagging a 1 m2 flannel cloth (Rulison et al., 2013; Ogden et al., 2018). Each grid comprised of a 7 x 7 array with 15 m between grid points. Each sample period comprised 8 parallel 90-m transects (alternating dragging and flagging), for a total of 720 m2. The cloth was inspected every 15 m. All ticks were removed and placed into vials of 95% ethanol. All three grids were sampled for a total of 2,160 m2 per sample period (also referred to as a ‘trapping session’ below). Wildlife hosts Sampling was carried out as described in Ginsberg et al., 2020 and Ogden et al., 2018. Each trapping session comprised of 2 trap nights. Small mammals were live captured from metal traps, pitfall traps, and occasionally wooden and metal coverboards. One long-folding aluminum Sherman trap (Sherman Traps, Tallahassee, FL) was placed at each of the 49 grid points (Appendix, Figure 3.10). Traps were baited with crimped oats. One pitfall array was placed external to and along each side of the grid. Each array comprised of two 40-m long aluminum flashing, buried into the soil, and bisecting each other perpendicularly. One five-gallon bucket was sunk into the ground at the end of each of the four arms, as well as the center of the array, for a total of five buckets per array and 20 buckets per grid. Twenty metal and twenty wooden (61 cm x 61 cm) coverboards were placed in pairs at 20 evenly spaced locations on the trapping array. To trap medium-sized mammals, one medium-sized Tomahawk trap (Tomahawk Live Trap Co. Tomahawk, WI) was placed at the midpoint of each edge of the grid, just outside of the grid for a 49 total of 4 traps. Traps were baited with one can of sardines. All traps were set in the evening and checked 12 hours later the next morning. Captured small mammals were identified to species, sexed, weighed, and marked with a uniquely numbered metal ear tag (Monel #1, National Band & Tag Company, Newport, KY). Medium-sized mammals were anesthetized and processed as described in Ogden et al., 2018. To minimize stress to the animal and to standardize sampling, we systematically inspected each animal for ticks for up to 5 minutes. All detected ticks were collected. We obtained ear biopsies (2 cm diameter) and blood samples (1% of body weight, up to 3 ml for medium-sized mammals) from each individual the first time we captured it within a trap session. If individuals were recaptured during the same trapping session, they were processed as described, but no additional ear biopsy nor blood sample was collected. Individuals recaptured in another trapping session were processed as if they were first time captures (i.e., with biopsy and blood samples collected). After animals were processed and were alert, we released them at the point of capture. Because of the lack of external ear pinnae, and because of their high metabolism, to minimize stress, we did not tag shrews individually and we did not sample blood. We prioritized sampling shrews first and then released them immediately before processing all other small mammals. All ticks and ear biopsies were stored in microcentrifuge tubes with 95% ethanol. We collected blood from hunter- harvested deer in Fort McCoy, Wisconsin. All animal handling procedures were approved by the Michigan State University's Institutional Animal Care and Use Committee (AUF # 06/09-094-00). Pathogen detection Questing ticks and ticks removed from hosts were identified at Michigan State University using dichotomous morphological keys (Clifford, Anastos and Elbl, 1961; Sonenshine, 1979; Keirans and Litwak, 1989; Durden and Keirans, 1997). For on-host ticks, to reduce the probability 50 that we would assay on-host larvae that had not fed long enough to acquire A. phagocytophilum from an infected host, we selected engorged ticks based on Han, 2019. Briefly, we used an 8-point scale Engorgement Index (EI) ranging from 1 (0-hour, attachment time point) to 8 (> 72 hours, natural drop-off time point). Up to 5 engorged larvae with EI7 (72 hours, right before natural drop- off time point) were selected by microscopic examination from each captured animal. Genomic DNA was extracted from questing ticks, ear biopsies, blood samples, and a subset of on-host I. scapularis ticks (both on-host nymphs and larvae based on the engorgement status), using Qiagen DNeasy Blood and Tissue kits (Valencia, CA) as per Hamer et al., 2010. The extracted DNA samples were assayed for A. phagocytophilum using a real-time PCR (rt-PCR) targeting a 122 bp region of the msp2 gene, which encodes a major surface protein (Drazenovich, Foley and Brown, 2006). A TaqMan probe-based rt-PCR on ABI QuantStudio 7 Flex PCR System was used at the Michigan State University Research Technology Support Facility Genomics Core. A total volume of 15 ml reaction consisted of 150 nM probe and each primer at 150 nM and 300 nM to assay tick DNA and mammalian tissues respectively. Bovine serum albumin at 0.25 ml was used as an additive in each reaction mixture to improve rt-PCR efficiency (Han, 2019). For the positive control, A. phagocytophilum human pathogenic Strain USG3 isolate derived from a beagle that was purposely exposed to infected adult I. scapularis ticks was used (Yeh et al., 1997) and was obtained from the Centers of Disease Control and Prevention (Yeh et al., 1997). For the negative control, ultrapure PCR water was used. Samples positive by the rt-PCR (i.e., “suspect positives”) were confirmed using a nested PCR that targets a 546 bp region of the 16S rRNA gene (Massung et al., 1998). The samples that were positive by the confirmatory nested PCR were sequenced by Sanger sequencing to distinguish between Ap-ha (human pathogenic strain) and Ap-v1 (deer variant strain) strains based on a two- 51 base pair difference (Massung et al., 2003) in the 16S rRNA gene. DNA sequencing was conducted by using the ABI Prism 7900HT Sequence Detection System and ABI 3730 xl DNA Analyzer (Applied Biosystems, Foster City, CA) at the Michigan State University Research Technology Support Facility Genomics Core. Definitions and data analysis Several objectives of this study include characterizing trends in A. phagocytophilum infection among questing I. scapularis, wildlife hosts, and transmission to on-host larval ticks. The prevalence of infection (%) of A. phagocytophilum in questing nymphs and adults was estimated by dividing the number of ticks confirmed A. phagocytophilum positive by rt-PCR divided by the total number of ticks tested for each respective life stage. Due to the relatively low infection prevalence of A. phagocytophilum in Wisconsin (on average 4-6% in nymphs (Lehane et al 2021; Foster et al. 2022) and 9-10% in adults (Foster et al. 2022; Lehane et al. 2021), to increase the reliability and representativeness in trends, we combined data from all grids and from all three years. Although all individuals were inspected for ticks each time they were captured, for ecological analysis, only the numbers of ticks collected during the first captures of individuals within a trapping period were used for estimating I. scapularis burden (i.e., number of ticks of a given life stage per individual). This is because for an individual captured for the first time within the trapping period, the number of ticks attached would comprise all ticks contacted for multiple days prior to capture. Because our protocol was to remove ticks from captured animals, however, the subsequent number of ticks on an individual recaptured the next day would not be comparable, as they would comprise newly parasitizing ticks and those that we missed on the first day of capture. Thus, for consistency of analyses, larval and nymphal burdens were estimated using only values 52 from an individual’s first capture within a trapping period. Because our trapping periods were 2- 3 weeks apart, a duration exceeding that for larval and nymphal I. scapularis to feed to repletion, data from animals recaptured in different trapping periods were included in all analyses. A negative binomial regression analysis was conducted to estimate if there were an effect of host species, sex, mass, trapping period (trapping week), trapping year and trapping array on the larval and nymphal counts on individual animals. Although A. phagocytophilum is an intracellular, blood-borne organism, researchers have used ear biopsy and spleen tissue biopsies to detect A. phagocytophilum in addition to using blood (Hodzic et al., 2001; Lane et al., 2005; Granquist, Aleksandersen, et al., 2010; Granquist, Stuen, et al., 2010; Szekeres et al., 2015; Rosso et al., 2017). Here we assayed both sample types (blood and ear biopsies), as each sample type was not always available from each animal or species (e.g., we did not collect blood samples from shrews). We scored any individual as positive (at a given time of capture) if at least one tissue type - blood or skin biopsy - was positive. The realized reservoir competence of a host species represents the probability that a larva feeding on an individual of a particular host species will become infected with A. phagocytophilum (Keesing et al., 2012). It was estimated as the mean percentage of on-host I. scapularis larvae infected by an individual host of a particular species, considering host infection prevalence (Keesing et al., 2012) (i.e., considering how not all individuals of a given host species are infected). Given equal sampling effort across sampling periods, the phenology, or seasonality, of questing ticks per month was estimated as the density of questing ticks (or questing infected ticks) detected in that month. Similarly, the phenology of on-host ticks (or infected on-host ticks) was calculated as the average number of larvae (or nymphs) per individual per host per month. We used a Chi-square test and Fisher's exact test (for n < 5 for any category) to compare the 53 infection prevalence between questing life stages and among hosts, applying a Bonferroni correction to the p-value when making multiple comparisons. We used logistic regression to compare the infection prevalence of A. phagocytophilum from biopsy samples and blood samples among species. A generalized linear model was used to model the A. phagocytophilum infection status of individual hosts as a function of host nymphal burden, host species, sampling grid, sampling year, trapping week, sex, and mass. Analyses were conducted using RStudio (Version 1.2.1335). All error bars shown for infection prevalence are 95% binomial confidence intervals. RESULTS The basic tick sampling and mammal trapping results were reported in Ogden et al. 2018, Ginsberg et al. 2020 and Han et al. 2019, but to understand A. phagocytophilum infection prevalence and dynamics, they are presented here as well. Furthermore, differences in presentations of results exist based on the different questions being asked. Questing ticks and pathogen detection Over three years and among all three grids, we collected 662 and 622 questing nymphal and adult I. scapularis, respectively. We found no significant differences in infection prevalence of questing nymphs and adults among the grids and sampling years, in part potentially due to low infection prevalence and sample sizes. Therefore, for subsequent analyses, the infection prevalence was combined overall years and grids. The overall infection prevalence for nymphs (15.3%, 12.6 - 18.2 95% CI) was lower than that of adults (20.4%, 17.3 - 23.8 95% CI, Fisher’s exact test p = 0.02301). Except for one nymph that was infected with the Ap-v1 variant, all other positive ticks (n=41 nymphs and n= 42 adults) were infected with the Ap-ha variant. All negative extractions and PCR controls were negative; all positive controls for PCR were positive. 54 Wildlife hosts Over three years, there were 2151 capture events comprising 14 small to medium-sized mammal species: white-footed mouse (Peromyscus leucopus), southern red-backed vole (Myodes gapperi), long-tailed shrew species (Sorex spp.), northern short-tailed shrew (Blarina brevicauda), raccoon (Procyon lotor), eastern chipmunk (Tamias striatus), meadow vole (Microtus pennsylvanicus), meadow jumping mouse (Zapus hudsonicus), Virginia opossum (Didelphis virginiana), star-nosed mole (Condylura cristata), southern flying squirrel (Glaucomys volans), fisher (Pekania pennanti), eastern mole (Scalopus aquaticus), and gray fox (Urocyon cinereoargenteus). Peromyscus leucopus and P. lotor were the most frequently captured small and medium-sized mammal species, respectively (Table 3.1). On-host ticks We collected a total of 8918 blacklegged ticks (8279 larvae, 623 nymphs, and 16 adults) from 2151 hosts (including recaptures), which comprised of 10 species (Table 3.1). When only the first-time captures within a trapping period were considered (including recaptures across trapping periods), we collected a total of 7204 larvae and 568 nymphs from 1822 individuals (Table 3.1). 55 Table 3.1. Small and medium-sized mammals captured at Fort McCoy, Wisconsin (2010 - 2012) and their average larval and nymphal I. scapularis burdens (number of ticks per individual host), median and range per host species. *First time captures could include recaptured individuals only if recaptures were caught during a different trapping period. Species First Tick infestation of Tick Burdens time first time captures captures within a trapping within a period trapping Total number of ticks Average number of period* collected (% ticks per individual prevalence) (median, range) Larvae Nymphs Larvae Nymphs White footed mouse 1055 5126 274 4.86 0.26 (Peromyscus leucopus) (73.3%) (16.6%) (3, 0-43) (0, 0-6) Southern red-backed vole 374 390 54 1.04 0.14 (Myodes gapperi) (40.6%) (10.2%) (0, 0-19) (0, 0-5) Long-tailed shrew (Sorex spp.) 164 626 8 4.07 0.05 (55.5%) (4.5%) (1, 0-42) (0, 0-2) Northern short-tail shrew 119 650 14 5.46 0.12 (Blarina brevicauda) (69.7%) (4.2%) (3, 0-43) ( 0, 0-4) Raccoon (Procyon lotor) 52 221 116 4.25 2.23 (38.5%) (55.8%) (0, 0-50) (1, 0-21) Eastern chipmunks (Tamias 27 84 71 3.1 2.63 striatus) (51.9%) (51.9%) (1, 0-21) (1, 0-10) Meadow vole 10 21 5 2.1 0.5 (Microtus pennsylvanicus) (60.0%) (20%) (1, 0-13) (0, 0-4) Meadow Jumping mouse 8 6 5 0.75 0 (Zapus hudsonius) (12.5%) (0%) (0, 0-6) Virginia opossum 7 22 22 3.14 3.14 (Didelphis virginiana) (42.9%) (51.1%) (0, 0-11) (0, 0-13) Star-nosed mole (Condylura 3 150 0 5 0 cristata) (66.70%) (0%) (7, 0-8) Southern flying squirrel 2 1 2 0.5 1 (Glaucomys volans) (50 %) (100%) (0.5, 0-1) (1, 1-1) Gray fox (Urocyon 1 0 2 0 2 cinereoargenteus) (0 %) (100%) (2, 0-2) 56 A negative binomial regression model indicated that host species, sex, mass, trapping period (trapping week), trapping year and trapping grid were significant factors in explaining the distribution of larval counts on individuals (Appendix, Table 3.5). The white-footed mouse, 2010, grid A, and males served as the reference levels. Southern red-backed vole had 0.19 times (95% confidence intervals (CIs) = 0.16 – 0.24, p-value = < 2e-16) fewer ticks, and the meadow vole had 0.35 times (95% CIs = 0.15 – 0.97, p-value = 0.03) fewer ticks compared to white-footed mice. Female mice had 0.65 times (95% CIs = 0.55 – 0.77, p-value = 3.70e-07) fewer larval counts compared to male mice. For each 1 g increase in mass, the larval count decreased by 1 (95% CIs 0.996 – 0.998, p-value=0.01). Larval counts per individual in 2011 were 0.52 times (95% CIs = 0.43 – 0.62, p-value = 9.96e-13) lower compared to 2010. With the progression of capture week over the season, the numbers of larvae parasitizing hosts decreased by 0.98 times per week (95% CIs 0.96 – 0.99, p-value = 0.0021). Individuals captured on grids B and C had 0.83 times (95% CIs = 0.69 – 0.99, p-value = 0.04) and 0.71 times (95% CIs = 0.59 – 0.87, p-value= 0.0006) fewer larvae compared to individuals captured on grid A (Appendix, Table 3.6). A negative binomial regression model indicated that host species, sex, mass, trapping period (trapping week), trapping year and trapping array were significant factors in explaining the distribution of nymphal counts on individuals (Appendix, Table 3.5). Again, the white-footed mouse, grid A, and male mice were used as the references levels. Compared to the white-footed mouse, the northern short-tail shrew, southern red-backed vole, and long-tailed shrew species had 0.10 times (95% CIs = 0.01 – 0.78, p-value = 0.03), 0.36 times (95% CIs = 0.25 – 0.53, p-value = 2.33e-07), and 0.24 times (95% CIs = 0.06 – 0.91, p-value = 0.04) fewer attached nymphs per individual. Conversely, the Virginia opossum, raccoon, and eastern chipmunk had 35 times (95% CIs = 9.1 – 132, p-value = 2.00e-07), 28 times (95% CIs = 8 – 101, p-value = 4.60e-07), and 8 57 times greater (95% CIs = 3.9 – 15.2, p-value = 5.09e-09) nymphs per individual respectively compared to the white-footed mouse. Compared to individuals captured on grid A, individuals captured on grid C had 0.6 times fewer nymphs (95% CIs = 0.42 – 0.87, p-value = 0.007). As the capture week progressed the number of nymphs on individual hosts significantly decreased by 0.89 times (95% CIs = 0.86 – 0.91, p-value = < 2e-16). With a 1 g increase in mass, the nymphal counts on hosts decreased by 1 time (95% CIs = 1.01 – 0.99, p-value = 0.03) (Appendix, Table 3.7). Pathogen detection Table 3.2 shows the pathogen detection results of 10 mammal species for which we had blood and/or biopsy samples. Individuals may have been captured in more than one trapping period, but blood (n = 516) and biopsy (n = 1212) samples were only collected once per trapping period. We were not able to collect blood or ear tissue species from every individual when captured (e.g., no blood was obtained from any shrew); hence the numbers of blood and ear biopsy samples are different. 58 Table 3.2. Anaplasma phagocytophilum infection of small to medium-sized mammals captured at Ft. McCoy, Wisconsin from 2010 – 2012 by sample type. *Includes multiple ear biopsies and blood samples from individuals captured from multiple trapping periods. ** The final number (%) of infected individuals comprised those from whom A. phagocytophilum had been detected from at least one tissue type. Host species Number of samples Number of host tissues positive assayed by tissue type (infection prevalence %) All Blood Ear All Blood* Ear tissues Biopsy tissues** Biopsy* 755 442 678 213 125 136 White-footed mouse (28.2) (28.3) (20.1) Southern red-backed 269 5 269 34 2 32 vole (12.6) (40.0) (11.9) Long-tailed shrews 75 0 75 5 - 5 spp. (6.7) (6.7) Northern short-tail 53 1 52 3 0 3 shrew (5.7) (0) (5.8) 51 40 47 17 11 10 Raccoon (33.3) (27.5) (21.3) 18 16 15 16 12 13 Eastern chipmunk (88.9) (75.0) (86.7) 9 1 9 2 0 2 Meadow vole (22.2) (0) (22.2) 7 6 7 4 2 2 Virginia opossum (57.1) (33.3) (33.3) Meadow jumping 7 1 7 1 1 0 mouse (14.3) (100.0) (0) Southern flying 2 0 2 0 - 0 squirrel (0) (0) 1 1 1 0 0 0 Star-nosed mole (0) (0) (0) Gray fox 1 1 1 0 0 0 (0) (0) (0) 59 Figure 3.1. Infection prevalence (95% binomial confidence intervals) of A. phagocytophilum for nine mammal species captured at Fort McCoy, Wisconsin from 2010 - 2012. Numbers of individuals assayed are shown beneath each bar. Anaplasma phagocytophilum was detected in 9 of the 12 mammal species captured (Table 3.2), and the overall infection prevalence of hosts captured was 23.6% (n=1248). Only Ap-ha strain was detected among infected individuals. Host species that had a greater sample size of 5 were used in subsequent analyses. The infection prevalence of eastern chipmunks (88.9%, n = 18) was highest and was significantly greater than that of the white-footed mouse, raccoon, southern red- backed vole, meadow vole, meadow jumping mouse, northern short-tailed shrew, and Sorex spp. (long-tailed shrews) (adjusted for multiple comparisons, pairwise Fisher's exact p-value < 0.001, Figure 3.1). There was no significant difference (adjusted for multiple comparisons, pairwise Fisher's exact p-value = 0.425) in the infection prevalence between the white-footed mice (28.0%, n = 762) and the raccoon (33.3%, n = 51). In addition, we screened 100 deer blood samples from 60 hunter-harvested deer. Only 3% of the deer were infected and all were infected with Ap-V1 strain. Host infectivity & realized reservoir competence We collected 7,146 larvae from nine host mammal species that were first-time captures within a trapping period, and which were of known sex (Table 3.1). White-footed mice (72%), northern short-tailed shrews (9%), long-tailed shrews spp. (9%), and southern red-backed voles (5%) comprised 95% of the attached larvae detected on captured hosts (Figure 3.2). Figure 3.2. Distribution of attached larvae detected among small to medium-sized mammalian species live captured at Fort McCoy, Wisconsin from 2010 – 2012. Error bars represent the 95% confidence intervals. Of the 7146 larvae, we selected 199 nearly fully engorged ticks to assay for A. 61 phagocytophilum (Table 3.4). These engorged larvae were collected predominantly from white- footed mice (78.4%) and eastern chipmunks (16.6%), from which A. phagocytophilum was detected in 18.6% (29/156) and 100% (33/33) of engorged larvae respectively. Given A. phagocytophilum was detected in 36% of white-footed mice and 81% of eastern chipmunks (Figure 3.1), the realized reservoir competence of these two host species is estimated as 12.9% (95% CIs 0.126 – 0.133) and 88.9% (95% CIs 0.861 – 0.813) respectively (Table 3.3). Although A. phagocytophilum was detected from many raccoons, southern red-backed voles, and shrews, no inferences about infectivity nor realized reservoir competence can be made as few or no engorged larvae were collected from these hosts that were infected with A. phagocytophilum. Table 3.3. The realized reservoir competence of white-footed mice and eastern chipmunks capture from 2010-2012 at Ft. McCoy, Wisconsin for Anaplasma phagocytophilum. Species Infection Total number Total Reservoir Realized prevalence (%) of fully number Competence* Reservoir (Total number engorged of larvae (SE) Competence** of captures larvae from infected (Confidence assayed) infected hosts with Ap Intervals) White- 28.2% 48 22 45.8% 12.9% footed (755) (9.3%) (0.126 – 0.133) mouse Eastern 88.9% 33 33 91.7% 88.9% chipmunk (18) (8.3%) (0.861 – 0.913) *Reservoir competence of a given species is calculated as the average infection prevalence of A. phagocytophilum in attached larvae sampled from A. phagocytophilum -infected hosts. 62 **Realized reservoir competence is calculated as the mean percentage of larvae infected with A. phagocytophilum by an individual host of a given species (Keesing et al., 2012), taking into consideration that not all individuals are infected. Factors affecting the infection status of a host A logistic regression model indicated that host species (Wald’s test p-value = 3e-09), B. burgdorferi infection status (Wald’s test p-value = 6.4e-05), trapping week (p-value = 0.000769), sex (Wald’s test p-value = 0.01), trapping grid (Wald’s test p-vale = 4.9e-07) and mass (Wald’s test p-value = 0.02) were significant factors for predicting A. phagocytophilum infection status (Appendix, Table 3.8). The nymphal burdens on hosts were not able to predict the variation in A. phagocytophilum status or the infection prevalence of host species (Figure 3.3, Appendix, Table 3.9) The odds of individuals being infected with A. phagocytophilum increased significantly by 2 times (95% CIs = 1.42 – 2.83, p-value = 8.62E-05) when individuals were infected with B. burgdorferi. Compared to white-footed mice, the odds of long-tailed shrew spp. being infected with A. phagocytophilum was 0.47 times less (95% CIs = 0.29 – 0.74, p-value = 0.001); the odds of meadow vole being infected was 15.3 times less (95% CIs = 1.75 – 173.5, p-value = 0.02); and the odds of the eastern chipmunk being infected was 23.44 times greater (95% CIs = 2.49 – 252.23, p-value = 0.005). The odds of females being infected with A. phagocytophilum was 1.5 times (95% CIs = 1.09 – 2.08, p-value = 0.013) lower compared to that of males. The odds of being infected with A. phagocytophilum decreased by 1 time (95% CIs = 1.00 – 0.99, p-value = 0.02) with a 1 g increase in mass (Appendix, Table 3.9). With the progression of capture week, the odds of being infected with A. phagocytophilum decreased by 0.95 times (95% CIs 0.92 – 0.98, p-value = 0.001). Compared to individuals captured on grid A, individuals trapped in grid B, the odds of being 63 infected were 2.89 times greater (95% CIs = 1.92 – 4.43, p-value = 5.93E-07) and for the individuals trapped on grid C, the odds of being infected increased by 1.5 times (95% CIs = 0.95 – 2.39, p-value = 5.93E-07). Figure 3.3. The relationship between the Anaplasma phagocytophilum infection prevalence and nymphal burdens among host species captured in Fort McCoy, Wisconsin from 2010 – 2012. The nymphal burden is the average number of nymphs per individual of a given host species. Enzootic cycle: phenology of interactions between ticks, white-footed mice, and A. phagocytophilum infections To examine the dynamics of infection between one generation of ticks to the next via a reservoir host, we focused on white-footed mice because they are considered an important reservoir host in the Northeast; there are an important host for juvenile blacklegged ticks at our field site, and because they formed the largest number of hosts we captured. Figure 3.4 depicts the phenology of questing I. scapularis ticks collected from Fort McCoy, Wisconsin averaged over the three study grids and years. The nymphal activity peaked in June and then gradually dropped 64 off as summer progressed, while the larval activity had a bimodal distribution with peaks in June and August (Figure 3.4). Figure 3.4. The phenology of questing I. scapularis ticks collected from Fort McCoy, Wisconsin from 2010 - 2012. The error bars show the 95% binomial confidence intervals. Similarly, we looked at the phenology of on-host I. scapularis ticks parasitizing all mammalian hosts (during their first-time captures within a trapping period) from May to October (Figure 3.5). Nymphs and larvae were observed on hosts in every month hosts were captured. The 65 phenology pattern was very similar to the questing tick phenology. On-host nymphal proportions showed an increase initially from May to July, but as the summer progressed, the proportions on- Figure 3.5. Phenology of I. scapularis parasitizing all mammal hosts captured at Ft. McCoy, Wisconsin, 2010-2012. The error bars represent the 95% binomial confidence intervals. host decreased by about 50%. Similar to the questing larvae, on-host larvae show a bimodal phenology, where the proportions of on-host larvae peaked in June and August. 66 We next looked at the phenology of the infection prevalence of A. phagocytophilum for all mammals together with the phenology of infection prevalence of A. phagocytophilum in on-host larvae parasitizing all hosts (Figure 3.6). Although not statistically significant, the trend in infection prevalence of A. phagocytophilum in all mammalian hosts increased from May – July and then decreased as the summer progressed. The infection prevalence of on-host larvae for A. phagocytophilum shows a similar pattern to that in mammalian hosts (Figure 3.6); although not statistically different, there is a trend that the larval peak occurs one month earlier. Figure 3.6. The phenology of infection prevalence of Anaplasma phagocytophilum in all captured mammals and on-host larvae and the density of questing infected nymphs (DIN) at Ft. McCoy, Wisconsin (2010-2012). The error bars represent the 95% binomial confidence intervals. Mouse blood v. biopsy v. transmission of A. phagocytophilum to on-host larvae In total, we assayed 1183 biopsy samples and 516 blood samples collected from 10 67 different host species (Table 3.4, Figure 3.7). The total host infection prevalence of A. phagocytophilum based upon all the biopsy samples was 18.7%, while based upon all blood samples, was 29.5% was significantly different (chi-square p-value < 0.00001). There were 427 animals from which both the blood and biopsy samples were taken during the same capture event. When considering only those host species that had at least 5 capture events, of which the white-footed mouse comprised the majority (Table 3.4), there was no significant difference in infection prevalence of A. phagocytophilum when detected by blood only, biopsy only and both types of tissues (Fisher’s exact test p-value ranged from 0.1 – 0.93) for each pairwise analyses. Table 3.4. Anaplasma phagocytophilum infection assay results from hosts with both blood and biopsy samples collected during the same capture event at Fort McCoy, Wisconsin from 2010 - 2012. Species listed had at least 5 capture events from different trapping periods. Host species Total Host tissues positive for A. phagocytophilum (infection number prevalence %) of capture Blood Biopsy only Both types of tissues (Blood events only and Biopsy) White-footed mouse 367 53 37 (10.1) 48 (13.1) (14.4) Raccoon 36 6 5 (13.9) 4 (11.1) (16.7) Eastern chipmunk 13 0 (0) 2 (15.5) 9 (69.2) Virginia opossum 6 2 2 (33.3) 0 (0) (33.3) Southern red-backed 5 2 0 (0) 0 (0) vole (40.0) 68 Because we had many A. phagocytophilum positives from blood and biopsy samples for white-footed mice (N=367), we conducted further analyses to compare blood and biopsy samples. Figure 3.7. Infection prevalence (95% binomial CI) of A. phagocytophilum in host animals where both a blood and biopsy tissue were sampled from the same capture event from Fort McCoy, Wisconsin from 2010 - 2012. N represents the total number of paired samples tested where some proportion of pairs tested positive for A. phagocytophilum in only the biopsy, in only the blood, or only in both the blood and biopsy. The infection prevalence of mice based on blood only, ear biopsies only, and both simultaneously were 14.4%, 10.1%, and 13.1% respectively. Overall, according to the pairwise Fisher’s exact test 69 with a Bonferroni correction there was no significant difference between detection of A. phagocytophilum among the sample types (p-values ranged between 0.08 – 0.2). We then compared the infection prevalence of mice that were positive by blood or biopsy regardless of if they were also positive by the other sample. In this case, again, there was no significant difference Figure 3.8. A Venn diagram of detections (%) of Anaplasma phagocytophilum from pairs of blood and biopsy samples collected from white-footed mice during the same capture event from Fort McCoy, Wisconsin (total N= 367 paired samples). In total, A. phagocytophilum was detected from 27.5% of blood (N = 53) and 23.2% of biopsy samples (N = 37). between the infection prevalence of mice when considering blood (27.5%; 95% CL 23.1 – 32.4%) or biopsies (23.2%; 95% CL 18.9 – 27.8%, Fisher’s exact value = 0.203 and p < 0.05). We next looked at how the infection prevalence of mice changed from May to October (compiled over all three years and three grids) (Figure 3.9). According to the pairwise Fisher’s 70 exact test with a Bonferroni correction, there was no significant difference in infection prevalence of A. phagocytophilum using blood only, biopsy only and both type of tissues simultaneously for each month (p-value ranged between 0.34 – 1). Figure 3.9. The infection prevalence of Anaplasma phagocytophilum in blood only, biopsy only and in both types of tissues simultaneously obtained from white-footed mice live captured at Fort McCoy, Wisconsin from 2010 - 2012. The error bars represent the 95% binomial confidence intervals. We conducted a logistic regression analysis to examine whether detection of A. phagocytophilum in the blood only, biopsy only, or both tissue types simultaneously, was related 71 to the B. burgdorferi infection status, the numbers of nymphs, sex, mass, trapping week and trapping array. In the first model, where A. phagocytophilum infection status by blood only was the response variable, capture year was a significant variable (Wald’s test p-value = 0.02), where the odds of being infected in 2012 was 2.69 times greater (95% CIs = 1.36 – 5.47, p-value = 0.005, Appendix, Table 3.10) compared to that in 2010. In the second model where A. phagocytophilum infection status by biopsy only was the response variable, the odds of being infected increased 1.08 times (95% CIs = 1.01 – 1.17, p-value = 0.034, Appendix, Table 3.11) with a 1 g increase in mass. In the third model where A. phagocytophilum infection status by both types of tissues simultaneously was the response variable, there was a significant difference in infection status among the capture years (Wald’s test p-value = 0.006), where compared to being captured in 2010, the odds of being infected in 2012 was lower by 0.23 times (95% CIs = 0.09 – 0.54, p-value = 0.001, Appendix, Table 3.12). With the progression of capture week, the odds of detection of A. phagocytophilum in both tissues simultaneously decreased by 0.92 times (95% CIs = 0.86 – 0.98, p-value = 0.02, Appendix, Table 3.12). No other factors significantly predicted infection status in each of the three models. We looked at the transmission of A. phagocytophilum from white-footed mice to larvae and how detection of Ap in each tissue type relates to the probability that a parasitizing larva would be infected. We developed a logistic regression model where the response variable in the model was the A. phagocytophilum infection status of a larva parasitizing an individual white-footed mouse, while the predictor variables were infection status for A. phagocytophilum by blood only, by biopsy only, or by both types of tissues simultaneously, the B. burgdorferi infection status of individuals, the number of parasitizing nymphs, sex, mass, trapping week and trapping array. In this model, the odds of detecting A. phagocytophilum in on-host larvae increased by 675 times 72 (95% CIs = 27 – 185660, p-value = 0.002, Appendix, Table 3.13) when A. phagocytophilum was detected in both types of tissues simultaneously compared to when A. phagocytophilum not detected only individuals. Also in this model, the odds of detecting A. phagocytophilum from on- host larvae increased by 1.47 times (95% CIs = 1.08 – 2.32, p-value = 0.04, Appendix, Table 3.13) with every 1 g increase of mass when A. phagocytophilum was detected in both types of tissues simultaneously. To try to infer the kinetics of infection from field data, we analyzed the temporal pattern of detecting A. phagocytophilum in individuals that were captured at least twice during the summer from different trapping periods, which were at least 2, if not 3, weeks apart. Appendix, Figure 3.11. shows the chronology of positive and negative assay results for A. phagocytophilum of individual mice when captured by Julian date. In total there were 40 mice that were captured multiple times that were positive for A. phagocytophilum by at least one type of tissue (blood, biopsy, or both) during at least one capture event (Appendix, Figure 3.11). As shown in Figure 3.11 there was no clear pattern in infection status of white-footed mice over time, but looking at general trends we noticed that animals become biopsy positive or remained biopsy positive at the end of summer (Panel A, Figure 3.11) between Julian days 250 – 300. Overall animals that become positive by blood only or remained blood positive and the animals that become positive by both tissues simultaneously and remained positive by both tissues simultaneously were found early to mid-summer (Panel B and Panel C, Figure 3.11) from Julian days 100 – 250. Most of the recaptured mice were not infected by both tissue types by the end of summer from Julian days 250 – 300. Early on in summer during the first 150 Julian days there were no animals that were infected by both types of tissues (blood and biopsy) at the same time. We also looked at how long A. phagocytophilum would be detected in an individual by 73 biopsy, blood and both types of tissue samples. On average, A. phagocytophilum was not detected after ~37- 38 days, with the range being ~16 – 93 days since first detection of infection. From the 40 individuals that were recaptured, individuals who were positive by both types of tissues simultaneously, we were able to detect A. phagocytophilum for about 18 days (16 – 93 days); for animals that were positive by biopsy only we were able to detect A. phagocytophilum for 74 days (range 33 – 93 days); and for animals who were positive only by blood we were able to detect A. phagocytophilum for 36 days (range 34 – 37 days). To try to infer transmission dynamics, using this same dataset, we used a logistic regression model to analyze if detection of an infected on-host engorged larva were related to trapping week, host infection status (0 = uninfected; 1 = biopsy only; 2 = blood only; 3 = blood and biopsy), sex, mass, nymphal burden, and trapping grid. In this model, the odds of larvae being infected with A. phagocytophilum significantly increased by 164.7 times (95% CIs value 17.6 – 4131.6, p-value = 1.4E-04) when mice were infected with A. phagocytophilum by both types of tissues simultaneously compared to individuals who were not infected, while the odds of larvae being infected with A. phagocytophilum significantly increased by 1.31 times (95% CIs value 1.07 – 1.70, p-value = 0.02, Appendix, Table 3.14) with every 1 g increase in mass. DISCUSSION The first detected A. phagocytophilum-infected patient in the US was reported from Wisconsin and Minnesota in the early 1990s (Chen et al., 1994), and retrospective analysis of adult I. scapularis ticks collected from northern Wisconsin in 1982 and 1991 revealed that 10.3% were infected with A. phagocytophilum (Pancholi et al. 1995). Anaplasma phagocytophilum had not been detected previously in North America and was first named Ehrlichia phagocytophila, and the disease was called human granulocytic ehrlichiosis (HGE) (Chen et al. 1994). The Upper Midwest, 74 namely Wisconsin and Minnesota, continues to be one of two major foci for human and canine anaplasmosis in the US. To better understand the ecology of the enzootic cycle of A. phagocytophilum in the Upper Midwest, we took advantage of an on-going field study by our lab to investigate the ecology of the Lyme disease bacterium at Fort McCoy, Monroe County, in the central western region of Wisconsin (Ogden et al., 2018; Ginsberg et al., 2020). An established population of I. scapularis has existed at Fort McCoy since at least since 1983 (Godsey et al., 1987), and several studies have demonstrated that B. burgdorferi, A. phagocytophilum, and other I. scapularis-borne pathogens are endemic at Ft. McCoy (Anderson, Duray and Magnarelli, 1987; Belongia et al., 1997; Steiner et al., 2008; Stromdahl et al., 2014; Han, Hickling and Tsao, 2016; Han et al., 2021). Prior investigations of A. phagocytophilum, however, either were focused only on ticks, or were conducted in the late 1990s, when A. phagocytophilum had not been detected and/or was present at below the level of detection in those studies. Here we had the opportunity to study the enzootic cycle where the prevalence of infection of A. phagocytophilum, and specifically that of the human pathogenic variant (Ap-ha), is quite robust. From a field study carried out over three years, we estimated the infection prevalence of A. phagocytophilum-ha in questing I. scapularis ticks; tick burdens on several host species; and compared the realized reservoir competence (Keesing et al., 2012, 2014) for A. phagocytophilum in several common small to medium-sized mammalian hosts that often-feed I. scapularis juvenile ticks. To infer the temporal dynamics of A. phagocytophilum transmission between I. scapularis and reservoir hosts, we characterized the phenology of A. phagocytophilum infection among questing nymphal I. scapularis, hosts, and attached larval I. scapularis. Finally, we also compared the detection of A. phagocytophilum between blood and ear tissue biopsies and related that to the detection of infected attached larval ticks. 75 Fort McCoy, an epidemiologically risky site for human granulocytic anaplasmosis The estimates for nymphal (15.3%, 12.6 - 18.2 95% CI) and adult (20.4%, 17.3 - 23.8 95% CI) infection prevalence for A. phagocytophilum observed at Ft. McCoy, Wisconsin from 2010 – 2012 are at the upper range of that previously reported in the literature for sites throughout Wisconsin (Pancholi et al. 1995; Lee et al., 2014; Murphy et al., 2017; Stauffer et al., 2020; Westwood et al., 2020), the Upper Midwest (Lehane et al., 2021; Burtis et al., 2022; Foster et al., 2022), as well as in the Northeast and Mid-Atlantic states (Lehane et al., 2021). Limited data from questing adult blacklegged ticks collected at Ft. McCoy suggests that A. phagocytophilum may have been emerging at Fort McCoy over the last two to three decades. Jackson et al., 2002 reported no detection from 713 adults I. scapularis ticks (nor 104 blood samples collected from white-footed mice) in 1997. The infection prevalence was 14% (n=100, 14% 95% CI 8.53-22.14) in adult ticks sampled in 2006 (Steiner et al., 2008) and 11.44% (n = 341, 8.48 – 15.42 95% CI) in 2006-2007 (Hamer et al., 2014). These data support the trend seen in Foster et al., 2022, which reports a marginally significant increase in adult infection prevalence for A. phagocytophilum at established sites that were sampled repeatedly from 2005-2019 in Minnesota and Wisconsin. Historically, most studies (Lehane et al., 2021) have not genetically differentiated A. phagocytophilum when detected in ticks or wildlife. As Massung et al., 1998 writes, however, lack of differentiation may be misleading since Ap-v1 has not been associated with human (nor canine) disease. In our study, apart from one nymph that harbored Ap-v1 (a deer variant strain), all infected questing nymphs and adults we collected harbored the Ap-ha variant. This finding agrees with that of the few prior investigations in Wisconsin where A. phagocytophilum has been typed (Steiner et al., 2008; Lee et al., 2014; Murphy et al., 2017), perhaps contributing to high anaplasmosis risk to 76 both humans and companion animals in this state. The prevalence of infection with Ap-ha in questing ticks appears to be more variable in the Northeast (Yeh et al., 1997; Courtney et al., 2003; Keesing et al., 2014; Edwards et al., 2019; Jordan, Gable and Egizi, 2022), where there are areas where Ap-v1 predominates (Massung et al., 1998; Edwards et al., 2019; Jordan, Gable and Egizi, 2022; Prusinski et al., 2023). More research is needed to understand under what conditions Ap-ha prevalence is high. One hypothesis is that Ap-v1 (and other variants) may compete with and reduce the prevalence of infection of Ap-ha, like that seen between different species of Rickettsia bacteria and Rickettsia rickettsii (Massung et al., 1998). Ixodes scapularis parasitizing wildlife at Fort McCoy, Wisconsin Ixodes scapularis is known to feed on many different host species, including mammals, birds, and lizards (Keirans et al., 1996). The juvenile stages of I. scapularis are responsible for the enzootic maintenance of non-vertically transmitted pathogens like B. burgdorferi and A. phagocytophilum. In our original study, Fort McCoy was one of eight field sites, where the objective was to compare the ecology of I. scapularis and B. burgdorferi over a latitudinal gradient from Wisconsin and Massachusetts to Florida. Specifically, the study was conducted to elucidate the tick-microbe-wildlife host interactions, focusing on the juvenile stages, as they are responsible for the enzootic maintenance of B. burgdorferi, which like A. phagocytophilum, is not vertically transmitted from adult females to larvae. Based on the known ecology at the time, limited resources, and capacity, we focused only on small and medium-sized mammals. Note, not reported in this dissertation, using the same PCR assays as reported here for Ft. McCoy, I detected 0% infection prevalence in both adults (N = 44) and nymphs (N = 9) at our field site in Tennessee and 4.9% in adults (N=142) and 2.3% (N = 256) in nymphs at our field site in Massachusetts respectively, supporting the clinal distribution of A. phagocytophilum observed by others (Lehane 77 et al. 2021). Using a variety of capture methods, over three years we captured 2151 hosts, comprising 14 species of small and medium-sized hosts. Most hosts comprised of small mammals, and of those, the white-footed mouse was frequently captured. The larval and nymphal burdens on captured hosts at our site resembled that previously reported where larvae most commonly are observed infesting small mammals such as white-footed mice, whereas nymphal life stages commonly infest larger hosts, including eastern chipmunks, raccoons and opossums (Anderson and Magnarelli, 1980; Davidar, Wilson and Ribeiro, 1989; Mather et al., 1989; Wilson et al., 1990; Shaw et al., 2003; Brunner and Ostfeld, 2008; Barbour et al., 2015; Jones et al., 2015). In particular, white- footed mice had the greatest average number of larvae per individual animal (4.86), while eastern chipmunks had the greatest average number of nymphs per individual (2.63), trends supported at other sites and studies (Mannelli et al., 1993; Slajchert et al., 1997; Schmidt, Ostfeld and Schauber, 1999; Shaw, Ostfeld and Keesing, 2001; Hamer et al., 2010; Han et al., 2021; Sidge et al., 2021) although differences between mouse and chipmunk larval loads may vary and may not be that different in some scenarios (e.g., Sidge et al. 2021, Keesing et al. 2009). Southern red-backed voles were the second most captured host species, but they had fewer on-host larvae and nymphs compared with white-footed, which has been seen in other studies (Main et al., 1982). Female hosts had significantly lower larval and nymphal infestations compared to males, which has been reported in other studies (Schmidt, Ostfeld and Schauber, 1999; Shaw, Ostfeld and Keesing, 2001; Brunner and Ostfeld, 2008; Jones et al., 2015). It has been hypothesized that males have more parasites on them on average, perhaps due to having relatively larger home ranges and having differences in reproduction and growth rates compared to females (Moore and Wilson, 2002; Krasnov et al., 2005; Butler et al., 2020). Temporally, although the three sampling grids 78 were similar in habitat and the number of questing larvae and questing nymphs collected were similar, but on-host larval and nymphal counts were lower in 2011. Seasonally, the phenological patterns of questing and on-host nymphs and larvae reflected the synchronous pattern, generally peaking in July, described for the Upper Midwest (Gatewood et al. 2009). Hosts infected with A. phagocytophilum in Fort McCoy, Wisconsin Because there is no transovarial transmission for A. phagocytophilum, its enzootic cycle is maintained through horizontal transmission. Naïve larvae may acquire A. phagocytophilum by feeding on an infected host. Alternatively, although there are contradictory results in the laboratory studies (Levin and Fish, 2000b, 2000a), field data suggest that blacklegged ticks may acquire A. phagocytophilum through non-systemic transmission by co-feeding in proximity with an infected nymph on the same host (Levin, Des Vignes and Fish, 1999; Levin and Fish, 2000b, 2000a). Conversely, the hosts can only become infected through the bite of an infected nymphal tick which obtained the pathogen as a larva. Thus, wildlife species that tend to be infested by both nymphs and larvae facilitate the maintenance of the enzootic cycle of A. phagocytophilum. At our study site the overall infection prevalence of mammals for A. phagocytophilum was about 24% (N = 1248). To our knowledge, this study is the first to investigate the infection prevalence of A. phagocytophilum among a community of small to medium sized mammals in the Midwest, and thus we cannot compare trends at the community level. If we focus on white-footed mice, our study found ~28% of the white-footed mice (n = 755) captured from 2010 – 2012 infected with the human pathogenic strain of A. phagocytophilum. Two studies conducted in the same region in Wisconsin but 10-14 years prior to ours found no white-footed mice infected with A. phagocytophilum. These studies were conducted at several study sites located near La Crosse, Wisconsin (Jackson et al., 2002) and within Fort McCoy, Wisconsin (Hofmeister et al., 1998). 79 These studies may represent a time just prior to the emergence of A. phagocytophilum. A more contemporaneous study conducted in northern Wisconsin from 2012 - 2014 found 1.7 % (n = 237) of the white-footed mice infected with A. phagocytophilum (Larson, Lee and Paskewitz, 2018); this difference may also reflect the emerging nature of A. phagocytophilum or ecological differences in host communities or climate influencing enzootic dynamics. A more recent study in north-central Wisconsin (2018-2019) showed that only 1% (n = 94) white-footed mice and 0.6% (n=318) deer mice (Peromyscus maniculatus), a closely related species, to be infected with A. phagocytophilum (Larson et al., 2021), As the authors point out, the first case of anaplasmosis from a human patient was detected in the county of their study, and thus in this case, it does not appear that A. phagocytophilum infection prevalence, at least in the small mammal community twenty years later, is very high. In Minnesota infections of A. phagocytophilum in white-footed mice varied from 11.4% (n = 158) in 1995 (Walls et al., 1997) to 46.8% (n=98) in 2000 and 20% (n=150) in 2001 at Camp Ripley, Minnesota (Johnson et al., 2011) Although these are only two studies, they may suggest that A. phagocytophilum became established earlier in eastern Minnesota. In the Northeast several studies have estimated A. phagocytophilum infection prevalence and reservoir competence within mammal communities and the infection prevalence ranged between 14.1% - 57.9% (Stafford et al., 1999; Levin, Nicholson, Massung and Fish, 2002; Massung et al., 2002; Keesing et al., 2012, 2014). Realized reservoir competence of a host species is estimated as the mean percentage of ticks infected by an individual of that host species (Keesing et al., 2012), and reflects on an animal’s ability to become infected by a pathogen and the ability to transmit the pathogen to a competent vector. According to Keesing et al., 2012, in a study conducted in southern New York, it was found that white-footed mice, eastern chipmunks, and short-tailed 80 shrews had greater realized reservoir competence for A. phagocytophilum (> 10%), compared to the other small to medium size mammals (as well as a few bird species captured at their study site in upstate New York). At our field site, the infection prevalence of the eastern chipmunk was significantly greater than any other small to medium-sized mammals captured at our field site and the eastern chipmunk had a higher realized reservoir competence compared to the white-footed mouse. In our study all engorged larvae that were tested from infected eastern chipmunks were infected with A. phagocytophilum, whereas not all engorged larvae collected from infected white- footed mice were infected. The realized reservoir competence values for both white-footed mouse and eastern chipmunk are comparable to those given in Keesing et al, 2012. All our host animals were infected with only Ap-ha which was slightly different from what Keesing et al, 2014, where most mammals species supported both strains, although most species had a higher reservoir competence for Ap-ha. A possible reason for this difference may be due in part to most questing ticks at our field site harboring mainly the Ap-ha strain as mentioned previously. The greater infection prevalence observed in eastern chipmunks compared with other species we captured and tested may be due to having on average higher nymphal burdens (Table 1, Figure 3). But interestingly nymphal burdens were not important in determining the infection status of A, phagocytophilum. Nymphal I. scapularis is the main life stage that is responsible for the transmission of A. phagocytophilum to small and medium-sized mammals. Having greater nymphal burdens ostensibly would result in a greater chance of A. phagocytophilum transmission to hosts. If the duration of detection of A. phagocytophilum in the blood and/or tissue were longer in eastern chipmunks, that might also result in having a greater infection prevalence compared to the white-footed mice. Although no laboratory transmission studies have been conducted with eastern chipmunks, in laboratory infected white-footed mice were able to launch an immune 81 response against A. phagocytophilum within 2 weeks of infection and antibodies remained in the blood for several months (Levin and Fish, 2000b). Furthermore, xenodiagnosis experiments using laboratory mice (Mus musculus) showed that majority of transmission to larvae occurred around 1-2 weeks post-infection of the mice although there was some variation among A. phagocytophilum isolates (Levin and Ross, 2004). Interestingly, there has been one study conducted on the congener redwood chipmunk (Tamias ochrogenys) in California, which was found to be PCR positive for A. phagocytophilum between four to seven weeks; and more importantly, even after 30 days post-infection was able to transmit A. phagocytophilum to xenodiagnostic larvae (Nieto and Foley, 2009). Thus, perhaps the duration of infection lasts longer in eastern chipmunks. Another hypothesis to explain the higher infection prevalence observed in eastern chipmunks is how the shorter life span and relatively higher reproductive rate of white- footed mice may dilute the infection prevalence over the season, especially if many offspring are born after the peak nymphal questing period. Even if eastern chipmunks were to have higher realized reservoir competence, however, white-footed mice may still play a greater role in terms of their relative contribution to the enzootic cycle of maintaining A. phagocytophilum. Given their high relative densities in habitats where I. scapularis is abundant, white-footed mice feed a relatively greater number of larvae compared to any other host species (Figure 2) and thus may transmit A. phagocytophilum to a greater number of larvae compared to other host species. In our study, we used multiple capture methods (Sherman live traps, pitfall traps, cover boards) with different capture biases, which may have allowed us to catch a greater diversity and number of small mammals compared to just using Sherman live traps. Amongst the species we captured, our relative numbers of animals caught per species suggest that the abundance of white-footed mice was likely greater compared to other host species captured. 82 With greater abundance and greater larval burdens, white-footed mice thus may contribute more to the maintenance of A. phagocytophilum compared to other hosts we trapped at our study site at Ft. McCoy, Wisconsin. Regarding the factors that might affect the infection status of a an individual, we found that that the B. burgdorferi infection status, host species, sex of the host species, mass of hosts species, trapping array and capture week were significant variables that explained the variation seen in the A. phagocytophilum infection status. In our model individuals who were infected with B. burgdorferi had two times the odds of being infected with A. phagocytophilum, which may not be surprising since both pathogens share the same vector and several reservoir host species. As seen by the estimates of infection prevalence, eastern chipmunks had greater odds of being infected with A. phagocytophilum compared to white-footed mice. Furthermore, in our comparison meadow vole and Sorex spp. had lower odds of becoming infected. Both these species, however, we did not have a larger sample size to determine accurately the infection prevalence. For meadow voles, the habitats we trapped may not have been optimal for them. In comparing the infection status between males and females, females had lower odds of being infected compared to males, for reasons discussed previously. A laboratory study reported that male laboratory mice had significantly higher loads of B. burgdorferi as well as higher prevalence in ear tissue (Zinck et al., 2022). The variation in A. phagocytophilum infection dynamics within hosts among different species and between sexes similarly should be further explored to understand factors influencing heterogeneities in transmission to larval I. scapularis. Phenology of ticks and infection of A. phagocytophilum To better understand the dynamics of A. phagocytophilum transmission between I. scapularis and wildlife hosts, we investigated the phenology of questing and on-host I. scapularis, 83 infection of A. phagocytophilum among the host community, and infection of A. phagocytophilum in on-host larvae. Phenological patterns of immatures are important to consider given the duration of infectivity of wildlife reservoir hosts for A. phagocytophilum, which is not vertically transmitted and for which some co-feeding, non-systemic transmission may occur. For A. phagocytophilum to be maintained enzootically, naïve larvae must acquire A. phagocytophilum from hosts that had been previously infected by an infected nymph (i.e., systemic transmission) or must co-feed near an infected nymph on the same host (i.e., non-systemic transmission). The shorter the duration of infectivity of the reservoir host, the more important the amount of overlap (or synchrony) between the nymphal and larval host-seeking activity periods. For instance, for the Lyme bacterium, B. burgdorferi, many epidemiologically important strains can infect white-footed mice for life. In this case, the duration of infectivity of a mouse is less important; larvae can host-seek later compared to nymphs; and factors such as survivorship and reproduction rates of the mouse population may be more important. At the other extreme, for tick-borne encephalitis (in Europe) and perhaps Powassan encephalitis virus (in North America), where co-feeding non-systemic transmission is believed to be the main route of horizontal transmission, enzootic maintenance necessitates great overlap of nymphal and larval host-seeking periods. A transmission experiment with laboratory mice (Mus musculus) showed that transmission to naive larvae (i.e., xenodiagnostic larvae) can occur for at least twelve weeks for some A. phagocytophilum isolates (Levin and Ross, 2004) but that higher transmission efficiency occurs 1-3 weeks post-infection, where 70-81% of xenodiagnostic larvae are infected. A second peak of transmission occurs about 3.5-7 weeks post-infection, when 36-64% of xenodiagnostic larvae become infected. In another experiment with laboratory bred white-footed mouse, Levin, and Fish (2000) presented data that suggest that mice can clear the infection by 11- and 15-weeks post- 84 infection. Furthermore, partial immunity reduces susceptibility to infection, transmission efficiency (in mice that do become infected) to xenodiagnostic larvae (~6-7% v. 83%) (Levin and Fish 2000), as well as non-systemic transmission. The percentage of larvae that acquired A. phagocytophilum from co-feeding infected nymphs was 10% on naïve mice and 1% on mice that had been infected previously (Levin and Fish, 2000a, 2000b). The phenologies of nymphal and larval host-seeking at Ft. McCoy resemble that previously observed for the Upper Midwestern region, where generally, there is broad overlap in the nymphal and larval host seeking phenologies (Gatewood et al., 2009; Hamer, Hickling, et al., 2012; Ogden et al., 2018). Given what is known about A. phagocytophilum infection and infectivity dynamics, the synchrony in nymphal and larval host-seeking phenologies at Ft. McCoy, and the Midwest in general, may facilitate both systemic and co-feeding transmission of A. phagocytophilum. This phenology of larvae differs from that typically found in the Northeast, where there is generally one peak of host-seeking larvae which occurs in late summer/early fall, about 2-3 months after the nymphal peak, and thus, where co-feeding transmission between nymphs and larvae may not occur as frequently (Gatewood et al., 2009; Hamer, Hickling, et al., 2012; Ogden et al., 2018). When we examine the phenology of the infection prevalence of A. phagocytophilum within the host community, it appears to lag by one month from the phenology of the density of questing infected nymphs and appears to match better the phenology of on-host nymphs. If we extrapolate from lab experiments with mice to the whole captured host community (which predominantly comprised white-footed mice), given peak infectivity of a host to larvae may occur 1-3 weeks post- exposure to an infected nymph, then the temporal trends in host infection prevalence generally suggest that at our study site, more than 60% of hosts are infected by July. After July, hosts still are acquiring infection, but at a lower rate, reflecting the decline in the density of questing infected 85 nymphs and on-host nymphs (Figure 6). A study conducted in Minnesota reported similar phenology in infected white-footed mice where the infection prevalence increased from May to June and then decreased as summer progressed (Johnson et al., 2011). Especially as we are uncertain about the relative abilities of hosts to transmit A. phagocytophilum based on detection of the DNA from blood, ear biopsy, and/or both tissue types, it is probably most informative to examine the phenology of acquisition of A. phagocytophilum by on-host larvae. Even though A. phagocytophilum-positive on-host larvae are detected May through October, a much larger proportion occur in the first half of the season - specifically, about 58 and 74% of all on-host larvae that test positive for A. phagocytophilum are collected by June and July respectively. This may reflect in part the phenology of on-host larvae, but interestingly the proportion of larvae in which A. phagocytophilum is detected appears to peak earlier than that in the host population. Though certainly not definitive, these data support the hypothesis that co- feeding transmission might contribute to the enzootic maintenance of A. phagocytophilum. The decline in infection in hosts and on-host larvae later in summer may reflect the clearance of A. phagocytophilum, reduced rate of exposure to infected nymphs (exacerbated by the birth of susceptible hosts), reduced susceptibility of hosts to re-infection and co-feeding transmission due to partial immunity (Levin and Fish 2000). In the Northeast, most larval activity peaks about 2.5 months after the nymphal activity peak. Thus, all else equal, phenological differences in I. scapularis activity between the upper Midwest and the Northeast should result in lower nymphal infection prevalence for A. phagocytophilum, but this has not been observed in limited data (Lehane et al 2021), and thus other factors (e.g., different strains, host differences in duration of infection, different host communities) may also be important. In study conducted between 2008 – 2012 the human granulocytic anaplasmosis case 86 incidence was higher in the Upper Midwest states compared to the Northeast (Dahlgren et al., 2015), but this no longer appears to be the case (Centers for Disease Control and Prevention, 2022). Thus, again, other factors, including those above as well as overall densities of ticks and differences in human behavior may be important. Biopsy vs Blood assays: What could we use to assay for A. phagocytophilum in field studies? Because A. phagocytophilum is a blood borne pathogen, for field studies involving wildlife, blood samples are generally collected. Our study determined that the animals frequently can be rt-PCR positive not just by blood but also by ear biopsy tissues. We compared assay results from matched ear biopsy and blood samples taken during the same capture event and found no significant difference in infection prevalence although pathogen detection results were not always congruent for samples taken at the same time. We then compared the infection prevalence by blood only, biopsy only and both tissues simultaneously for a few hosts species and within white-footed mice. Again, there was no significant difference in infection prevalence of A. phagocytophilum. We also did not see any difference in the phenology of infection prevalence by blood, biopsy or both types of tissues. We next considered just white-footed mice since we had relatively large sample size of matched individuals, and conducted a logistic regression analysis to see if the host factors or environmental factors would affect the likelihood that an individual would be infected by blood only or biopsy only or both types of tissues simultaneously. In each model, we had different factors that came out as significant, but it is unclear what these factors reveal about the infection dynamics of A. phagocytophilum by blood, biopsy or both types of tissues simultaneously. When considering transmission of A. phagocytophilum to larvae from white-footed mice we wanted to see if there were any association with the detection of A. phagocytophilum by tissue 87 type. Our model showed that when an individual mouse was infected with both types of tissues simultaneously, the odds of detecting A. phagocytophilum in on-host larvae were 675 times greater (95% CIs = 27 – 185660, p-value = 0.002) than if A. phagocytophilum were detected in only the blood or the biopsy. It could mean that when A. phagocytophilum is detected in both types of tissues at the same time, there may be a higher load of bacteria within the hosts and therefore the probability of transmission is higher. The best method to further examine these factors would be to conduct a xenodiagnostic lab experiment like that of Levin and Fish 2000 and Levin and Ross et al. 2004, where one would infect white-footed mice with A. phagocytophilum and simultaneously assay biopsy, blood, and transmission efficacy to larval I. scapularis for at least 16 weeks, a time frame that would simulate natural enzootic dynamics in the Upper Midwest. Several studies in Europe have considered using ear biopsy samples over blood samples because A. phagocytophilum might be short-lived in the blood (J S Liz et al., 2000; Kevin J. Bown et al., 2003; Bown et al., 2008; Baráková et al., 2014; Rosso et al., 2017), which might explain why there might be relatively fewer animals infected by blood only later in the season. To test this hypothesis, conducting a laboratory transmission experiment using white-footed mice and measuring the length of detection of A. phagocytophilum by blood and biopsy would be useful. A study conducted on dogs found that even after dogs seroconverted, A. phagocytophilum could be detected in ear biopsies (Berzina et al., 2014), and it was hypothesized in separate study that the skin may be a site for persistent infection of A. phagocytophilum (Ladbury et al., 2008; Granquist, Aleksandersen, et al., 2010) which could be the reason for our observations in this study. But currently no study exists to explain how A. phagocytophilum can persist in the skin. Anaplasma phagocytophilum is the second most common vector-borne disease in the United States (Adams et al., 2017). The human case incidence of granulocytic anaplasmosis in the 88 US drastically increased from ~ 340 cases in 2000 to over 5600 cases in 2019 (Biggs, Behravesh, K. Bradley, et al., 2016; Adams et al., 2017; Dumic et al., 2022b). With the increase in and expansion of I. scapularis populations, the human disease incidence of granulocytic anaplasmosis will increase, therefore investigating the ecology of A. phagocytophilum should help in prevention of human granulocytic anaplasmosis (e.g., how well would bait tubes designed to kill I. scapularis on mice reduce anaplasmosis risk to humans?). Our investigation focused on examining dynamics over three years at one site where I. scapularis and associated diseases have been endemic. Given the variation in climate and vegetation throughout the Midwest, more ecological studies should be conducted to better understand factors affecting the heterogeneity in the enzootic cycles leading to variation in spatial risk of anaplasmosis. Future studies should include bird species known to harbor I. scapularis ticks, especially to better appreciate the potential of birds for dispersing infected I. scapularis ticks into emerging regions. Additional studies examining strain diversity and wildlife host species in the Midwest may help to further understand heterogeneity of A. phagocytophilum risk and predict risk as it continues to emerge throughout the eastern U.S. 89 BIBLIOGRAPHY Adams, D.A. et al. (2015) Summary of Notifiable Infectious Diseases and Conditions — United States, 2015. Available at: https://www.cdc.gov/MMWR/. Anderson, J.F., Duray, P.H. and Magnarelli, L.A. (1987) ‘Prevalence of Borrelia burdorferi in white-footed mice and Ixodes dammini at Fort McCoy, Wis.’, Journal of Clinical Microbiology, 25(8), pp. 1495–1497. Available at: https://doi.org/10.1128/jcm.25.8.1495-1497.1987. Anderson, J.F. and Magnarellp, L.A. (1980) ‘Vertebrate host relationships and distribution of Ixodid ticks (Acari: Ixodidae)’, J. Med. Entomol, pp. 314–323. Available at: https://doi.org/10.1093/jmedent/17.4.314. Arsnoe, I.M. et al. (2015) ‘Different populations of blacklegged tick nymphs exhibit differences in questing behavior that have implications for human lyme disease risk’, PLoS ONE, 10(5), pp. 1–21. Available at: https://doi.org/10.1371/journal.pone.0127450. Baráková, I. et al. (2014) ‘Genetic and ecologic variability among Anaplasma phagocytophilum strains, Northern Italy’, Emerging Infectious Diseases, 20(6), pp. 1082–1085. Available at: https://doi.org/10.3201/eid2006.131023. Barbour, A.G. et al. (2015) ‘Association between body size and reservoir competence of mammals bearing Borrelia burgdorferi at an endemic site in the northeastern United States’, Parasites and Vectors, 8(1), pp. 1–5. Available at: https://doi.org/10.1186/s13071-015-0903-5. Barlough, J.E. et al. (1997) ‘Ehrlichia phagocytophila genogroup rickettsiae in ixodid ticks from California collected in 1995 and 1996’, Journal of Clinical Microbiology, 35(8), pp. 2018–2021. Available at: https://doi.org/10.1128/jcm.35.8.2018-2021.1997. Belongia, E.A. et al. (1997) ‘Prevalence of granulocytic Ehrlichia infection among white-tailed deer in Wisconsin’, Journal of Clinical Microbiology, 35(6), pp. 1465–1468. Available at: https://doi.org/10.1128/jcm.35.6.1465-1468.1997. Berzina, I. et al. (2014) ‘Anaplasma phagocytophilum DNA amplified from lesional skin of seropositive dogs’, Ticks and Tick-borne Diseases, 5(3), pp. 329–335. Available at: https://doi.org/10.1016/j.ttbdis.2013.12.010. Biggs, H.M. et al. (2016) ‘Diagnosis and management of tickborne rickettsial diseases: Rocky Mountain spotted fever and other spotted fever group rickettsioses, ehrlichioses, and anaplasmosis -United States A Practical Guide for Health Care and Public Health Professionals’. Centers for Disease Control and Prevention. Available at: http://www.cdc.gov/mmwr/cme/conted.html. Bown, K.J. et al. (2003) ‘Seasonal dynamics of Anaplasma phagocytophila in a rodent-tick (Ixodes trianguliceps) system, United Kingdom’, Emerging Infectious Diseases, 9(1), pp. 63–70. Available at: https://doi.org/10.3201/eid0901.020169. 90 Bown, K.J. et al. (2008) ‘Relative importance of Ixodes ricinus and Ixodes trianguliceps as vectors for Anaplasma phagocytophilum and Babesia microti in field vole (Microtus agrestis) populations’, Applied and Environmental Microbiology, 74(23), pp. 7118–7125. Available at: https://doi.org/10.1128/AEM.00625-08. Brunner, J.L. and Ostfeld, R.S. (2008) ‘Multiple causes of variable tick burdens on small- mammal hosts’, Ecology, 89(8), pp. 2259–2272. Available at: https://doi.org/10.1890/07-0665.1. Burtis, J.C. et al. (2022) ‘Predicting distributions of blacklegged ticks (Ixodes scapularis), Lyme disease spirochetes (Borrelia burgdorferi sensu stricto) and human Lyme disease cases in the eastern United States’, Ticks and Tick-borne Diseases, 13(5). Available at: https://doi.org/10.1016/j.ttbdis.2022.102000. Butler, R.A. et al. (2020) ‘Small-mammal characteristics affect tick communities in southwestern Tennessee (USA)’, International Journal for Parasitology: Parasites and Wildlife, 12, pp. 150–154. Available at: https://doi.org/10.1016/j.ijppaw.2020.05.012. Chen, S.M. et al. (1994) ‘Identification of a granulocytotropic Ehrlichia species as the etiologic agent of human disease.’, Journal of Clinical Microbiology, 32(3), pp. 589–595. Available at: https://doi.org/10.1128/jcm.32.3.589-595.1994. Clifford, C.M., Anastos, G. and Elbl, A. (1961) ‘The larval Ixodid ticks of the Eastern United States (Acarina - Ixodidae)’, Misc. Publ. of the Entomological Society of America, 2(5), pp. 213– 237. Courtney, J.W. et al. (2003) ‘Molecular characterization of Anaplasma phagocytophilum and Borrelia burgdorferi in Ixodes scapularis ticks from Pennsylvania’, Journal of Clinical Microbiology, 41(4), pp. 1569–1573. Available at: https://doi.org/10.1128/JCM.41.4.1569- 1573.2003. Dahlgren, F.S. et al. (2015) ‘Human granulocytic anaplasmosis in the United States from 2008 to 2012: a summary of national surveillance data.’, The American journal of tropical medicine and hygiene, 93(1), pp. 66–72. Available at: https://doi.org/10.4269/ajtmh.15-0122. Davidar, P., Wilson, M. and Ribeiro, J.M.C. (1989) ‘differential distribution of immature Ixodes dammini (acari: Ixodidae) on rodent hosts’, Journal of Parasitology, 75(6), pp. 898–904. Available at: https://doi.org/10.2307/3282868. Drazenovich, N., Foley, J. and Brown, R.N. (2006) ‘Use of Real-time quantitative PCR targeting the msp2 protein gene to identify cryptic Anaplasma phagocytophilum infections in wildlife and domestic animals’, Vector Borne and Zoonotic Disease, 6(1), pp. 83–90. Available at: https://doi.org/10.1089/ast.2007.0153. Dugat, T. et al. (2015) ‘Opening the black box of Anaplasma phagocytophilum diversity: current situation and future perspectives.’, Frontiers in cellular and infection microbiology, 5, pp. 1–18. 91 Available at: https://doi.org/10.3389/fcimb.2015.00061. Dumic, I. et al. (2022) ‘Human Granulocytic Anaplasmosis—a systematic review of published Cases’, Microorganisms. MDPI. Available at: https://doi.org/10.3390/microorganisms10071433. Durden, L.A. and Keirans, J.E. (1997) ‘Nymphs of the genus Ixodes (Acari: Ixodidae) of the United States: taxonomy, identification key, distribution, hosts, and medical/veterinary importance’, Thomas Say Publications in Entomology: Monographs. Entomological Society of America (Thomas Say publications in entomology: Monographs). Available at: https://doi.org/10.2307/3495570. Edwards, M.J. et al. (2019) ‘A 4-Yr survey of the range of ticks and tick-Borne pathogens in the Lehigh Valley region of eastern Pennsylvania’, Journal of Medical Entomology, 56(4), pp. 1122–1134. Available at: https://doi.org/10.1093/jme/tjz043. Foster, E. et al. (2022) ‘Inter-annual variation in prevalence of Borrelia burgdorferi sensu stricto and Anaplasma phagocytophilum in host-seeking Ixodes scapularis (Acari: Ixodidae) at long- term surveillance sites in the upper midwestern United States: Implications for public hea’, Ticks and Tick-borne Diseases, 13(2), p. 101886. Available at: https://doi.org/10.1016/j.ttbdis.2021.101886. Gatewood, A.G. et al. (2009) ‘Climate and tick seasonality are predictors of Borrelia burgdorferi genotype distribution’, Applied and Environmental Microbiology, 75(8), pp. 2476–2483. Available at: https://doi.org/10.1128/AEM.02633-08. Ginsberg, H.S. et al. (2020) ‘Local abundance of Ixodes scapularis in forests: effects of environmental moisture, vegetation characteristics, and host abundance’, Ticks and Tick-borne Diseases, 11(1), p. 101271. Available at: https://doi.org/10.1016/j.ttbdis.2019.101271. Ginsberg, H.S. et al. (2021) ‘Why Lyme disease is common in the northern US, but rare in the south: The roles of host choice, host-seeking behavior, and tick density’, PLOS Biology, 19(1), p. e3001066. Available at: https://doi.org/10.1371/journal.pbio.3001066. Godsey, M.S. et al. (1987) ‘Lyme disease ecology in Wisconsin: distribution and host preferences of Ixodes dammini, and prevalence of antibody to Borrelia burgdorferi in small mammals’, American Journal of Tropical Medicine and Hygiene, 37(1), pp. 180–187. Available at: https://doi.org/10.4269/ajtmh.1987.37.180. Granquist, E.G., Aleksandersen, M., et al. (2010) ‘A morphological and molecular study of Anaplasma phagocytophilum transmission events at the time of Ixodes ricinus tick bite.’, Acta veterinaria Scandinavica, 52(1), p. 43. Available at: https://doi.org/10.1186/1751-0147-52-43. Granquist, E.G., Stuen, S., et al. (2010) ‘Variant-specific and diminishing immune responses towards the highly variable MSP2(P44) outer membrane protein of Anaplasma phagocytophilum during persistent infection in lambs’, Veterinary Immunology and Immunopathology, 133(2–4), pp. 117–124. Available at: https://doi.org/10.1016/j.vetimm.2009.07.009. 92 Grassi, L. et al. (2021) ‘Ecotyping of Anaplasma Phagocytophilum from wild ungulates and ticks shows circulation of zoonotic strains in northeastern Italy’, Animals, 11(2), pp. 1–14. Available at: https://doi.org/10.3390/ani11020310. Hamer, S.A. et al. (2010) ‘Invasion of the Lyme disease vector Ixodes scapularis: implications for Borrelia burgdorferi endemicity’, EcoHealth, 7(1), pp. 47–63. Available at: https://doi.org/10.1007/s10393-010-0287-0. Hamer, S.A. et al. (2012) ‘Synchronous phenology of juvenile Ixodes scapularis, vertebrate host relationships, and associated patterns of Borrelia burgdorferi ribotypes in the midwestern United States’, Ticks and Tick-borne Diseases, 3(2), pp. 65–74. Available at: https://doi.org/10.1016/j.ttbdis.2011.11.004. Hamer, S.A. et al. (2014) ‘Increased diversity of zoonotic pathogens and Borrelia burgdorferi strains in established versus incipient Ixodes scapularis populations across the Midwestern United States’, Infection, Genetics and Evolution, 27, pp. 531–542. Available at: https://doi.org/10.1016/j.meegid.2014.06.003. Han, S. (2019) ‘How is Borrelia miyamotoi maintained among its vector, Ixodes scapularis and vertebrate host population?’ Dissertation. Michigan State University. Available at: https://doi.org/https://doi.org/doi:10.25335/M56M3376J. Han, S. et al. (2021) ‘Seasonality of acarological risk of exposure to Borrelia miyamotoi from questing life stages of Ixodes scapularis collected from Wisconsin and Massachusetts, USA’, Ticks and Tick-borne Diseases, 12(1), p. 101556. Available at: https://doi.org/10.1016/j.ttbdis.2020.101556. Han, S., Hickling, G.J. and Tsao, J.I. (2016) ‘High prevalence of Borrelia miyamotoi among adult blacklegged ticks from white-tailed deer’, Emerging Infectious Diseases, 22(2), pp. 316– 318. Available at: https://doi.org/10.3201/eid2202.151218. Hodzic, E. et al. (2001) ‘Acquisition Dynamics of Borrelia burgdorferi and the Agent of Human Granulocytic Ehrlichiosis at the Host–Vector Interface’, \Vector Borne and Zoonotic Diseases, 1(2), pp. 107–115. Available at: https://doi.org/10.4324/9781003232506-9. Hofmeister, E.K. et al. (1998) ‘Cosegregation of a novel Bartonella species with Borrelia burgdorferi and Babesia microti in Peromyscus leucopus’, The Journal of Infectious Diseases, 177(2), pp. 409–416. Available at: https://doi.org/https://doi.org/10.1086/514201. Jackson, C.A. et al. (2002) ‘Reassessment of a midwestern Lyme disease focus for Borrelia burgdorferi and the human granulocytic ehrlichiosis agent’, Journal of Clinical Microbiology, 40(6), pp. 2070–2073. Available at: https://doi.org/10.1128/JCM.40.6.2070-2073.2002. Johnson, R.C. et al. (2011) ‘Agents of human anaplasmosis and Lyme Disease at Camp Ripley, Minnesota’, Vector-Borne and Zoonotic Diseases, 11(12), pp. 1529–1534. Available at: 93 https://doi.org/10.1089/vbz.2011.0633. Jones, C.R. et al. (2015) ‘Factors affecting larval tick feeding success: Host, density and time’, Parasites and Vectors, 8(1). Available at: https://doi.org/10.1186/s13071-015-0955-6. Jordan, R.A., Gable, S. and Egizi, A. (2022) ‘Relevance of spatial and temporal trends in nymphal tick density and infection prevalence for public health and surveillance practice in long- term endemic areas: a case study in Monmouth County, NJ’, Journal of Medical Entomology, 59(4), pp. 1451–1466. Available at: https://doi.org/10.1093/jme/tjac073. Keesing, F. et al. (2012) ‘Reservoir competence of vertebrate hosts for Anaplasma phagocytophilum’, Emerging Infectious Diseases, 18(12), pp. 10–13. Available at: https://doi.org/10.3201/eid1812.120919. Keesing, F. et al. (2014) ‘Prevalence of human-Active and variant 1 strains of the tick-borne pathogen Anaplasma phagocytophilum in hosts and forests of Eastern North America’, American Journal of Tropical Medicine and Hygiene, 91(2), pp. 302–309. Available at: https://doi.org/10.4269/ajtmh.13-0525. Keirans, J.E. et al. (1996) ‘Ixodes scapularis (Acari: Ixodidae): redescription of all active stages, distribution, hosts, geographical variation, and medical and veterinary importance’, Journal of Medical Entomology, 33(3), pp. 297–318. Available at: https://doi.org/10.1093/jmedent/33.3.297. Keirans, J.E. and Litwak, T.R. (1989) ‘Pictorial key to the ddults of hard ticks, family Ixodidae (Ixodida: Ixodoidea), east of the Mississippi River’, Journal of medical entomology, 26(5), pp. 345–448. Available at: https://doi.org/10.1093/jmedent/26.5.435. Krasnov, B.R. et al. (2005) ‘Sex-biased parasitism, seasonality and sexual size dimorphism in desert rodents’, Oecologia, 146(2), pp. 209–217. Available at: https://doi.org/10.1007/s00442- 005-0189-y. Ladbury, G.A.F. et al. (2008) ‘Dynamic transmission of numerous Anaplasma phagocytophilum genotypes among lambs in an infected sheep flock in an area of anaplasmosis endemicity’, Journal of Clinical Microbiology, 46(5), pp. 1686–1691. Available at: https://doi.org/10.1128/JCM.02068-07. Lane, R.S. et al. (2005) ‘Western gray squirrel (Rodentia:Sciuridae) : A primary reservoir host of Borrelia burgdorferi in California oak woodlands?’, Journal of Medical Entomology, 42(4), pp. 388–396. Available at: https://doi.org/https://doi.org/10.1093/jmedent/42.3.388. Larson, S.R., Lee, X. and Paskewitz, S.M. (2018) ‘Prevalence of tick-borne pathogens in two species of peromyscus mice common in Northern Wisconsin’, Journal of Medical Entomology, 55(4), pp. 1002–1010. Available at: https://doi.org/10.1093/jme/tjy027. Lee, X. et al. (2014) ‘Prevalence of Borrelia burgdorferi and Anaplasma phagocytophilum in 94 Ixodes scapularis (Acari: Ixodidae) nymphs collected in managed red pine forests in Wisconsin.’, Journal of medical entomology, 51(3), pp. 694–701. Available at: https://doi.org/10.1603/ME13140. Lehane, A. et al. (2021) ‘Prevalence of single and coinfections of human pathogens in Ixodes ticks from five geographical regions in the United States, 2013–2019’, Ticks and Tick-borne Diseases, 12(2), p. 101637. Available at: https://doi.org/10.1016/j.ttbdis.2020.101637. Levin, M.L. et al. (2002) ‘Comparison of reservoir competence of medium sized mammals and Peromyscus leucopus for Anaplasma phagocytophilum in Connecticut’, Vector Borne and Zoonotic Disease, 2(3). Available at: https://doi.org/10.1089/ast.2007.0153. Levin, M.L. and Fish, D. (2000) ‘Immunity reduces reservoir host competence of Peromyscus leucopus for Ehrlichia phagocytophila’, Infection and Immunity, 68(3), pp. 1514–1518. Available at: https://doi.org/10.1128/IAI.68.3.1514-1518.2000. Levin, M.L. and Ross, D.E. (2004) ‘Acquisition of different isolates of Anaplasma phagocytophilum by Ixodes scapularis from a model animal.’, Vector-borne and zoonotic diseases, 4(1), pp. 53–59. Available at: https://doi.org/10.1089/153036604773082997. Liveris, D. et al. (2021) ‘A new genetic approach to distinguish strains of Anaplasma phagocytophilum that appear not to cause human disease’, Ticks and Tick-borne Diseases, 12(3). Available at: https://doi.org/10.1016/j.ttbdis.2021.101659. Liz, J S et al. (2000) ‘PCR detection of granulocytic ehrlichiae in Ixodes ricinus ticks and wild small mammals in western Switzerland.’, Journal of clinical microbiology, 38(3), pp. 1002–7. Available at: https://doi.org/10.1128/JCM.38.3.1002-1007.2000. Main, A.J. et al. (1982) ‘Immature Ixodes dammini (Acari:Ixodid) on small animals in Connecticut, USA’, J. Med. Entomol, 19(6), pp. 655–664. Available at: https://academic.oup.com/jme/article/19/6/655/2220015. Majazki, J. et al. (2013) ‘Anaplasma phagocytophilum strains from voles and shrews exhibit specific ankA gene sequences.’, BMC veterinary research, 9, p. 235. Available at: https://doi.org/10.1186/1746-6148-9-235. Mannelli, A. et al. (1993) ‘Ixodes dammini (Acari: Ixodidae) Infestation on Medium-Sized Mammals and Blue Jays in Northwestern Illinois’, J. Med. Entomol. Available at: https://academic.oup.com/jme/article/30/5/950/2221321. Marumoto, K. et al. (2007) ‘Detection of Anaplasma phagocytophilum and Ehrlichia sp. HF strains in Ixodes ricinus ticks in Brittany, France’, Clinical Microbiology and Infection, 13(3), pp. 338–341. Available at: https://doi.org/10.1111/j.1469-0691.2006.01630.x. Massung, R.F. et al. (1998) ‘Nested PCR Assay for Detection of Granulocytic Ehrlichiae’, Journal of Clinical Microbiology, 36(4), pp. 1090–1095. Available at: 95 https://doi.org/10.1128/jcm.36.4.1090-1095.1998. Massung, R.F. et al. (2002) ‘Genetic variants of Ehrlichia phagocytophila, Rhode Island and Connecticut’, Emerging Infectious Diseases, 8(5), pp. 467–472. Available at: https://doi.org/10.3201/eid0805.010251. Massung, R.F. et al. (2003) ‘Inability of a variant strain of Anaplasma phagocytophilum to infect mice.’, The Journal of infectious diseases, 188(11), pp. 1757–63. Available at: https://doi.org/10.1086/379725. Mather, T.N. et al. (1989) ‘Comparting the relative potential of rodents as reservoirs of the Lyme disease spirochete (Borrelia burgdorferi)’, American journal of epidemiology. Available at: https://doi.org/https://doi.org/10.1093/oxfordjournals.aje.a115306. Moore, S.L. and Wilson, K. (2002) ‘Parasites as a viability cost of sexual selection in natural populations of mammals’, Science, 297(5589), pp. 2015–2018. Available at: https://doi.org/10.1126/science.1074196. Murphy, D.S. et al. (2017) ‘Prevalence and Distribution of Human and Tick Infections with the Ehrlichia muris -Like Agent and Anaplasma phagocytophilum in Wisconsin, 2009–2015’, Vector-Borne and Zoonotic Diseases, 17(4), pp. 229–236. Available at: https://doi.org/10.1089/vbz.2016.2055. Nieto, N.C. and Foley, J.E. (2009) ‘Reservoir Competence of the Redwood Chipmunk ( Tamias Ochrogenys ) for Anaplasma Phagocytophilum’, Vector-Borne and Zoonotic Diseases, 9(6), pp. 573–577. Available at: https://doi.org/10.1089/vbz.2008.0142. Ogden, N.H. et al. (2018) ‘Evidence for geographic variation in life-cycle processes affecting - phenology of the Lyme disease vector Ixodes scapularis (Acari: Ixodidae) in the United States’, Journal of Medical Entomology, 55(6), pp. 1386–1401. Available at: https://doi.org/10.1093/jme/tjy104. Pancholi, P. et al. (1995) ‘Ixodes dammini as a potential vector of human granulocytic ehrlichiosis’, The Journal of Infectious Diseases, 172(4), pp. 1007–1012. Available at: https://doi.org/10.1093/infdis/172.4.1007. Prusinski, M. et al. (2023) ‘Associations of Anaplasma phagocytophilum bacteria variants in Ixodes scapularis ticks and humans, New York, USA’, Emerging Infectious Diseases, 29(3), pp. 540–550. Available at: https://doi.org/10.3201/eid2903.220320. Rar, V. and Golovljova, I. (2011) ‘Anaplasma, Ehrlichia, and “ Candidatus Neoehrlichia” bacteria: Pathogenicity, biodiversity, and molecular genetic characteristics, a review’, Infection, Genetics and Evolution, 11(8), pp. 1842–1861. Available at: https://doi.org/10.1016/j.meegid.2011.09.019. Rejmanek, D. et al. (2013) ‘Unique strains of Anaplasma phagocytophilum segregate among 96 diverse questing and non-questing Ixodes tick species in the western United States’, Ticks and Tick-borne Diseases, 4(6), pp. 482–487. Available at: https://doi.org/10.1016/j.ttbdis.2013.06.003. Richter Jr., P.J. et al. (1996) ‘Ixodes pacificus (Acari: Ixodidae) as a Vector of Ehrlichia equi (Rickettsiales: Ehrlichieae)’, Journal of Medical Entomology, 33(1), pp. 1–5. Available at: https://doi.org/10.1093/jmedent/33.1.1. Rosso, F. et al. (2017) ‘Prevalence and genetic variability of Anaplasma phagocytophilum in wild rodents from the Italian alps’, Parasites & Vectors, 10(1), p. 293. Available at: https://doi.org/10.1186/s13071-017-2221-6. Rulison, E.L. et al. (2013) ‘Flagging Versus Dragging as Sampling Methods for Nymphal Ixodes scapularis (Acari : Ixodidae ) Flagging versus dragging as sampling methods for nymphal Ixodes scapularis ( Acari : Ixodidae )’, Journal of Vector Ecology, 38(1), pp. 163–167. Available at: https://doi.org/10.1111/j.1948-7134.2013.12022.x. Schmidt, K.A., Ostfeld, R.S. and Schauber, E.M. (1999) ‘Infestation of Peromyscus leucopus and Tamias striatus by Ixodes scapularis (Acari: Ixodidae) in Relation to the Abundance of Hosts and Parasites’, J. Med. Entomol. Available at: https://academic.oup.com/jme/article/36/6/749/903627. Shaw, M., Ostfeld, S. and Keesing, F. (2001) ‘Infestation of Peromyscus leucopus and Tamias striatus by Ixodes scapularis (Acari: Ixodidae) as a result of tick host preference, grooming efficiency and habitat utilization’, Undergraduate Ecology Research Reports. Millbrook, NY: Institute of Ecosystem Studies., pp. 1–15. Shaw, M.T. et al. (2003) ‘Factors influencing the distribution of larval blacklegged ticks on rodent hosts’, American Journal of Tropical Medicine and Hygiene, pp. 447–452. Available at: https://doi.org/10.4269/ajtmh.2003.68.447. Sidge, J.L. et al. (2021) ‘Lake Michigan insights from island studies: the roles of chipmunks and coyotes in maintaining Ixodes scapularis and Borrelia burgdorferi in the absence of white-tailed deer’, Ticks and Tick-borne Diseases, 12(5). Available at: https://doi.org/10.1016/j.ttbdis.2021.101761. Silaghi, C. et al. (2017) ‘Guidelines for the direct detection of Anaplasma spp. in diagnosis and epidemiological studies’, Vector-Borne and Zoonotic Diseases, 17(1), pp. 12–22. Available at: https://doi.org/10.1089/vbz.2016.1960. Slajchert, T. et al. (1997) ‘Role of the eastern chipmunk (Tamias Striatus) in the epizootiology of Lyme borreliosis in Northwestern Illinois, Usa’, Journal of Wildlife Diseases, 33(1), pp. 40–46. Available at: https://doi.org/10.7589/0090-3558-33.1.40. Sonenshine, D.E. (1979) Insects of Virginia, Research Division Bulletin. Virginia poyltechnic institute and state university. 97 Stafford, K.C. et al. (1999) ‘Infection with agents of human granulocytic ehrlichiosis, Lyme disease, and babesiosis in wild white-footed mice (Peromyscus leucopus) in Connecticut’, 37(9), pp. 2887–2892. Available at: https://doi.org/10.1128/jcm.37.9.2887-2892.1999. Stauffer, M.T. et al. (2020) ‘Detection of zoonotic human pathogens from Ixodes scapularis in Wisconsin’, Journal of Vector Ecology, 45(1), pp. 147–149. Available at: https://doi.org/10.1111/jvec.12384. Steiner, F.E. et al. (2008) ‘Infection and co-infection rates of Anaplasma phagocytophilum variants, Babesia spp., Borrelia burgdorferi, and the rickettsial endosymbiont in Ixodes scapularis (Acari: Ixodidae) from sites in Indiana, Maine, Pennsylvania, and Wisconsin’, Journal of Medical Entomology, 45(2), pp. 289–297. Available at: https://doi.org/10.1603/0022- 2585(2008)45[289:IACROA]2.0.CO;2. Stromdahl, E. et al. (2014) ‘Comparison of phenology and pathogen prevalence, including infection with the Ehrlichia muris-like (EML) agent, of Ixodes scapularis removed from soldiers in the midwestern and the northeastern United States over a 15-year period (1997-2012)’, Parasites and Vectors, 7(1), pp. 1–12. Available at: https://doi.org/10.1186/s13071-014-0553-z. Stuen, S. (2007) ‘Anaplasma Phagocytophilum - The most widespread tick-borne infection in animals in Europe’, Veterinary Research Communications, 31(SUPPL. 1), pp. 79–84. Available at: https://doi.org/10.1007/s11259-007-0071-y. Stuen, S., Granquist, E.G. and Silaghi, C. (2013) ‘Anaplasma phagocytophilum--a widespread multi-host pathogen with highly adaptive strategies.’, Frontiers in cellular and infection microbiology, 3(July), p. 31. Available at: https://doi.org/10.3389/fcimb.2013.00031. Szekeres, S. et al. (2015) ‘Candidatus Neoehrlichia mikurensis and Anaplasma phagocytophilum in natural rodent and tick communities in Southern Hungary’, Ticks and Tick-borne Diseases, 6(2), pp. 111–116. Available at: https://doi.org/10.1016/j.ttbdis.2014.10.004. Telford, S.R. et al. (1996) ‘Perpetuation of the agent of human granulocytic ehrlichiosis in a deer tick-rodent cycle’, Proceedings of the National Academy of Sciences of the United States of America, 93(12), pp. 6209–6214. Available at: https://doi.org/10.1073/pnas.93.12.6209. Tveten, A.-K. (2014) ‘Prevalence and diversity among Anaplasma phagocytophilum Strains originating from Ixodes ricinus ticks from northwest Norway ’, Journal of Pathogens, 2014, pp. 1–8. Available at: https://doi.org/10.1155/2014/824897. Walls, J.J. et al. (1997) ‘Natural infection of small mammal species in Minnesota with the agent of human granulocytic ehrlichiosis.’, Journal of clinical microbiology, 35(4), pp. 853–5. Available at: https://doi.org/10.1128/jcm.35.4.853-855.1997. Westwood, M.L., Peters, J.L. and Rooney, T.P. (2020) ‘Prevalence and coinfection of three tick- borne pathogens in questing adult blacklegged ticks Ixodes scapularis (Vilas County, 98 Wisconsin)’, Vector-Borne and Zoonotic Diseases, 20(8), pp. 633–635. Available at: https://doi.org/10.1089/vbz.2020.2619. Wilson, M.L. et al. (1990) Host-dependent differences in feeding and reproduction of Ixodes dammini (Acari: Ixodidae), J. Med. Entomol, pp. 945–954. Available at: https://doi.org/https://doi.org/10.1093/jmedent/27.6.945. Yeh, M.T. et al. (1997) ‘Serologic and molecular detection of granulocytic ehrlichiosis in Rhode Island’, Journal of Clinical Microbiology, 35(4), pp. 944–947. Available at: https://doi.org/10.1128/jcm.35.4.944-947.1997. 99 APPENDIX NSF-EID ARRAY LAYOUT March 17, 2010 b a c e 5 5 A-PD2 6 6 7 7 8 8 d A- A-Track A-Tom3 A-Track A- Cam2 2 3 Cam3 (A1-7) (A2-7) (A3-7) (A4-7) (A5-7) (A6-7) (A7-7) 4 4 X X X X X X X 9 9 b (A1-6) (A2-6) (A3-6) (A4-6) (A5-6) (A6-6) (A7-6) X X X X X X X a c e A-PD3 d 17 17 18 18 (A1-5) (A2-5) (A3-5) (A4-5) (A5-5) (A6-5) (A7-5) 3 3 X X X X X X X 10 10 A-Tom2 (A1-4) (A2-4) (A3-4) (A4-4) (A5-4) (A6-4) (A7-4) X X X X X X X (A1-3) (A2-3) (A3-3) (A4-3) (A5-3) (A6-3) (A7-3) 2 2 X X X X X X X 11 11 A-Tom3 19 19 20 20 b (A1-2) (A2-2) (A3-2) APPENDIX (A4-2) (A5-2) (A6-2) (A7-2) a c X X X X X X X e A-PD1 d A-MC1 (A2-1) (A3-1) (A4-1) (A5-1) (A6-1) (A7-1) (A1-1) X X X X X X X 12 12 A-WC1 A- A-Track A-Track A- Cam1 1 A-Tom4 4 Cam4 16 16 b 15 15 14 14 13 13 a c e KEY: Metal cover board (MC) A-PD4 d Wooden cover board (WC) (A1-2) Sherman trap X A- Cam4 Camera 0 10 20 30 A-Tom4 Tomahawk trap Scale (m) A-Track Track Plate 1 Pitfall/drift- fence array (PD) Figure 3.10. The set-up of traps within a sampling grid at Fort McCoy, Wisconsin. 100 Figure 3.11. Anaplasma phagocytophilum infection status through time in individual recaptured white-footed mouse (rows) at Fort McCoy, Wisconsin 2010-2012. Each panel represents the infection status of an individual mouse by a particular tissue type: biopsy only (A), blood only (B) and by both tissues simultaneously (C). “●” represents animals from which A. phagocytophilum was not detected in that tissue type, while a “▲” represents animals which A. phagocytophilum was detected in that tissue type. For reference Julian date 121, 152, 182, 213, 244, 274 and 305 represents May 1st, June 1st, July 1st, August 1st, September 1st, October 1st, and November 1st. 101 Figure 3.11 (cont’d) 102 Table 3.5. The negative binomial regression model coefficients and the incident rate ratios for the Ixodes scapularis larvae parasitizing small to medium sized mammals. Independent Variables Coefficients Std Error Z-value p-value Incidence rate ratio (95% CIs) Intercept 2.36 0.13 18.07 < 2e-16 10.53(8.07 - 13.81) Compared to white-footed mice Northern short-tail shrew -0.02 0.23 -0.09 0.93 0.98 (0.64 - 1.56) Southern red-backed vole -1.65 0.10 -15.71 < 2e-16 0.19 (0.16 - 0.24) Virginia opossum -0.05 0.63 -0.08 0.94 0.95 (0.32 - 3.86) Meadow vole -1.05 0.47 -2.21 0.03 0.35 (0.15 - 0.97) Raccoon 0.94 0.57 1.65 0.10 2.55 (1.07 - 6.94) Sorex spp. -0.30 0.26 -1.17 0.24 0.74 0.45 - 1.27) Eastern chipmunks -0.42 0.31 -1.36 0.18 0.66 (0.37 - 1.25) Meadow Jumping mouse -1.89 0.67 -2.80 0.00 0.15 (0.04 - 0.60) Compared to sampling year 2010 2011 -0.66 0.09 -7.13 9.96e-13 0.51 (0.43 - 0.62) 2012 -0.03 0.09 -0.35 0.73 0.97 (0.82 - 1.15) Capture week -0.02 0.01 -3.21 0.001 0.98 (0.96 - 0.99) Compared to sampling grid A Grid B -0.19 0.09 -2.09 0.04 0.83 (0.69 - 0.99) Grid C -0.34 0.10 -3.44 0.0006 0.71 (0.59 - 0.87) Compared to males Female -0.43 0.09 -5.08 3.70e-07 0.65 (0.55 - 0.77) Mass -0.00021 0.00 -2.55 0.010821 1.00 (0.9997 - 0.9999) 103 Table 3.6. The negative binomial regression model coefficients and the incident rate ratios for the Ixodes scapularis nymphs parasitizing small to medium sized mammals. Independent Variables Coefficients Std Error Z-value p-value Incidence rate ratio (95% CIs) Intercept 0.11 0.23 0.50 0.62 1.12 (0.72 - 1.76) Compared to white-footed mice Northern short-tail shrew -2.29 1.04 -2.20 0.03 0.1 (0.01 - 0.78) Southern red-backed vole -1.01 0.20 -5.17 2.33e-07 0.36 (0.25 - 0.53) Virginia opossum 3.55 0.68 5.20 2.00e-07 34.66 (9.11 - 131.93) Meadow vole 0.48 0.71 0.67 0.50 1.61 (0.40 - 6.50) Raccoon 3.32 0.66 5.04 4.60e-07 27.73 (7.62 - 100.91) Sorex spp. -1.44 0.69 -2.09 0.04 0.24 (0.06 - 0.91) Eastern chipmunks 2.04 0.35 5.84 5.09e-09 7.66 (3.87 - 15.17) Meadow Jumping mouse -29.03 893500 0.00 1 2.98E-13 (0.00 – Inf) Compared to sampling year 2010 2011 0.69 0.17 4.18 2.94e-05 1.99 (1.44 - 2.75) 2012 2.08 0.17 1.25 0.21 1.23 (0.89 - 1.70) Capture week -0.12 0.01 -9.084 < 2e-16 0.89 (0.86 - 0.91) Compared to sampling grid A Grid B -0.20 0.16 -1.24 0.22 0.81 (0.59 - 1.11) Grid C -0.50 0.18 -2.70 0.007 0.61 (0.42 - 0.87) Compared to males Female -0.59 0.17 -3.53 0.0004 0.56 (0.4 - 0.77) Mass -0.00021 0.00010 -2.12 0.03 1.00 (1.01 - 0.999) 104 Table 3.7. The Wald's test p-values for the effect of each categorical variable on the Anaplasma phagocytophilum infection status of host mammals. Variable Wald’s test p-value Effect of species 3.00E-09 Effect of capture year 0.12 Effect of sex 0.01 Effect of grids 4.90E-07 105 Table 3.8. The logistical regression model coefficients and the odd ratios for the Anaplasma phagocytophilum infection status of hosts mammals. Independent Variable Coefficients Std Error Z-value p-value Odds Ratio Intercept -1.27 0.32 -5.24 1.57E-07 0.19 (0.10 - .035) Borrelia burgdorferi infection status 0.69 0.18 3.93 8.62E-05 2.00 (1.42 - 2.83) Compared to white-footed mice Northern short-tail shrew -0.77 1.13 -0.69 0.492 0.46 (0.02 - 3.08) Southern red-backed vole 15.47 392.10 0.04 0.969 5213287 (4.6e-6 - Inf) Virginia opossum -1.27 1.10 -1.16 0.246 0.28 (0.01 - 1.67) Meadow vole 2.73 1.15 2.38 0.017 15.34 (1.75 - 173.51) Raccoon -0.34 0.66 -0.511 0.609 0.71 (0.16 - 2.32) Sorex spp -0.76 0.24 -3.19 0.001 0.5 (0.29 - 0.74) Eastern chipmunks 3.15 1.13 2.79 0.005 23.43 (2.50 - 252.23) Meadow Jumping mouse -1.01 1.10 -0.92 0.359 0.37 (0.02 - 2.21) Number of nymphs parasitizing individual hosts 0.15 0.08 1.74 0.082 1.16 (0.98 - 1.37) Compared to sampling year 2010 2011 0.39 0.22 1.74 0.082 1.47 (0.95 - 2.28) 2012 -0.08 0.19 -0.42 0.678 0.92 (0.64 - 1.33) Capture week -0.05 0.02 -3.36 0.001 0.95 (0.92 - 0.98) Compared to sampling grid A Grid B 1.06 0.21 4.99 5.93E-07 2.89 (1.92 - 4.43) Grid C 0.41 0.24 1.72 0.085 1.50 (0.95 - 2.39) Compared to males Male 0.40830 0.16380 2.493 0.013 1.50 (1.09 - 2.08) Mass -3.99E-04 1.71E-04 -2.33 0.020 1.00 (1 - 0.999) 106 CHAPTER 4: SPECIES DISTRIBUTION MODELING AND APPLICATION TO IXODES SCAPULARIS ABSTRACT Species distribution modeling (SDM) is a tool to identify environmental predictors important for a species and to predict suitable habitats and therefore spatial distribution of a species. Here I present a limited review of SDM approaches and then review SDMs of Ixodes scapularis, the tick vector that is responsible for many vector-borne diseases in the U.S. There are several methods of developing SDM’s, which overall are based on either regression-based approaches or machine learning approaches. Each method incorporates presence-absence or abundance data and is a based on different sets of assumptions. With the development of many species distribution modeling techniques, many have been applied to vector borne disease systems to identify potential environmental variables and habitats that are important to sustaining vectors and therefore predict where vectors and their associated diseases may spread in the future. Developing SDMs enable researchers and public health agencies to strategize better surveillance efforts, especially given limited resources. Species distribution models can be used to inform the public and healthcare workers about the risk in contacting a vector and associated pathogen(s) in regions where the vector previously was not known to occur or be established. These models can also be important when developing control and prevention plans. Developing SDMs for suitable habitats for vectors were first attempted in the early 1970s for mosquitoes, and since then there have been a marked improvement in using SDMs to model potential habitats for mosquitoes especially in the developing world. With the increase in ticks and tickborne disease in the US, due in part to expansion of geographic ranges of several vector species, SDM’s have been applied to predict the spread of several tick species. The blacklegged tick, Ixodes scapularis, is responsible for transmitting the 107 most common vector borne disease in the US, Lyme disease. With land use changes, forest management, wildlife management, and climate change, I. scapularis has been expanding its range for several decades. In the late 1990s, when I. scapularis populations in Wisconsin were still limited geographically, the first I. scapularis SDM was developed. Since then, several SDMs for I. scapularis have been developed using different types of data, different spatial scales, and different modeling approaches. Keywords: Species distribution models, vector borne diseases, Ixodes scapularis, range expansion 108 INTRODUCTION Species Distribution Modeling Species distribution modeling (SDM) has become one of the most important modeling approaches in ecology. These models incorporate ecological concepts with those of the natural history of a species using statistical methods (Elith and Leathwick, 2009). The origin of SDMs stems from early studies based on mapping out species distribution across landscape along with their associations with geographic and environmental features (Elith and Leathwick, 2009). Many early studies used a combination of climatic features with environmental elements to explain different vegetation patterns seen across the world. These models could be applied to species from any taxa, both fauna and flora from marine to terrestrial ecosystems. Species distribution models can provide insight into where suitable habitats lie for a species of interest, such as one that is endangered, or alternatively one that is invasive. SDMs typically consider the known localities for a species, which could be occurrence or abundance data, and then try to identify environmental predictors which are associated, and therefore may influence, the distribution of the species of interest (Duarte, Whitlock and Peterson, 2018). The niche concept of a species forms the central theorem for SDMs. In terms of SDMs, a species’ niche is defined as the place an organism exists within its environment or community with all the abiotic and biotic interactions it has (Grinnell, 1917; Whittaker, Levin and Root, 1973). For SDMs, the niche is refined further into a fundamental niche, which encompasses the broad range of environmental features an organism can survive in, and the realized niche, which is the actual environment the organism occupies within the fundamental niche due to different environmental constraints (Whittaker, Levin and Root, 1973; Duarte, Whitlock and Peterson, 2018). A SDM will 109 essentially predict a species’ ability to survive within an environmental space using a statistical method that will quantitatively describe the environmental space (fundamental niche) that is suitability for that species to survive in (Duarte, Whitlock and Peterson, 2018). The final species distribution model will display suitable habitats within a geographic extent that will have all the estimated suitable environmental conditions conducive for the species to survive and reproduce. Response and Predictor Variables in SDMs A critical component for building a SDM is to measure all the known occurrence points of a species of interest within a geographic extent. Essentially species occurrence points could be measured as presence/absence of that species or abundance. The level of measurement will depend upon what data are available and which model type is used. The next critical component that is needed for model development are the predictor variables, i.e., the environmental variables that will be used to estimate environmental suitability. Typically, climatic, habitat and soil variables are the most used predictors. The next step is to determine the modeling algorithm to use. Species distribution modeling algorithms Species distribution modeling algorithms fall into two main categories: 1) regression-based methods and 2) machine learning based methods. I will briefly review them here. 1. Regression-based methods Regression-based models like generalized linear models (GLMs), and its non-parametric form a generalized additive model (GAM), are straightforward and the most frequently used approaches. Depending on what type of variable is the response variable, the form of the GLM would change. For example, if the response variable is presence and absence points of a species, a logistical regression model is used, while if the response variable is counts of individuals of a species, a Poisson regression may be used. Other types of linear regression may be used depending 110 on the distribution of counts (e.g., negative binomial or zero-inflated negative binomial). Generalized additive models closely resemble GLMs but tend to be more flexible because the model is fit using smoothing and piecewise linear splines (Duarte, Whitlock and Peterson, 2018). In general, a GAM has lower bias than a GLM but tends to have a higher variance because the coefficients (bs) of GLM are replaced by a smoothing spline in GAMs (Elith and Franklin, 2013). A more advanced form of a GAM currently being used is a multivariate adaptive regression spline (MARS), which is computationally faster than a GAM and the predictions can be easily converted into a geographical information system (GIS) mapping format (Duarte, Whitlock and Peterson, 2018). But MARs can only be used when the response variable data are converted into a normally distributed dataset, not when the responses are at a presence/absence data level. 2. Machine learning-based methods A relatively new development in SDMs is the use of machine learning and data mining techniques to predict the distribution of species. These methods include A) boosted regression trees; B) random forests; C) artificial neural networks; D) genetic algorithms; and E) maximum entropy methods. A. Boosted Regression Trees Boosted regression trees will combine several models using two algorithms, a decision or regression tree and a boosting regression (Elith, Leathwick and Hastie, 2008). In a decision tree, a set of predictor variables are used, and the input data are split repeatedly according to the predictor variables (Elith, Leathwick and Hastie, 2008). The decision to split the data at the node is to increase the information gained from the tree. The splitting of data will continue until a stopping condition is put into the model. The decision trees on its own is not very accurate because the tree will depend on the input sample data and the features. A slight change in the sample data can result 111 in a completely different sequences of node splits which might not reflect on the actual relationship. Therefore, each decision tree is boosted to improve the accuracy in this model. Boosted regression trees will build several simple iterative decision trees to the training dataset and combine these trees to give a more accurate representation between the distribution of the species and the environmental variables (Li and Wang, 2013). B. Random Forests Random forests are like boosted regression trees in that this technique uses many iterative decision trees which are then bootstrapped to find the optimal model (Drew, Wiersma and Huettmann, 2011). Random forests are based on the Breiman’s random forest algorithm, where random decision trees are built depending on the training dataset. The final model will be chosen by the number of votes or classifications each of the trees gets in the random forest; the tree which has the highest vote will be the final model (Breiman, 2001; Drew, Wiersma and Huettmann, 2011). Both the boosted regression trees and random forests are based on ensemble methods where several models or trees are combined to obtain the best model to explain the distribution of the species. Both these techniques are powerful computational machine learning methods. C. Artificial Neural Networks Artificial neural networks are another complex modeling method which involve a network of simple processing elements or artificial neurons (Li and Wang, 2013). This method looks at the links between the environment layers and the species distribution dataset which acts as neurons. The artificial neural networks contain hidden layers with neurons receiving information from the input data which are then summed up and added with a constant, which usually is a bias. These then are transformed using a fixed function (Li and Wang, 2013; Zhang and Li, 2018). 112 D. Genetic Algorithms Genetic algorithms are based on presence only data of a species and use preset mathematical rules which are called genes. These rules are combined randomly to develop models to explain patterns of distribution for a species (Stockwell and Peters, 1999; Li and Wang, 2013). E. Maximum Entropy The maximum entropy (Maxent) method is also a presence only data model. The concept behind the Maxent method is to estimate a target probability distribution by finding the probability distribution that is at a maximum entropy (either spread out or closest to a uniform distribution) that is subjected to a set of environmental constraints (Phillips, Anderson and Schapire, 2006). Each method has its own set of advantages and disadvantages. The most important aspect is choosing the method that seems most appropriate to predict the distribution of a species of interest. The decision to use a specific method of SDMs will depend mainly on the input sampling data set. For example, if a study has conducted a passive surveillance, there is only information about presence points; therefore, using a machine learning technique which only uses presence points is more appropriate. Once the model is developed, model evaluation is conducted, if possible, to estimate the accuracy of the developed distribution map. The most common way to evaluate a model is assessing the area under the curve (AUC) of the receiver operating curve (ROC) area. The AUC is based on the sensitivity (when the model predicts presence points accurately, i.e. true positives) and specificity (when model predicts absence points accurately i.e., true negatives) of the model data, where the data are split into a training set to develop the model, and the testing data set to test how well the model performs (Fielding and Bell, 1997a). The AUC values range from 0 – 1, where when the AUC values are greater than 0.5, the model will be able to classify suitable habitats better than by chance or at random (Jiménez-Valverde, 2012). Using the AUC to evaluate 113 a model is beneficial when we have known absence points for a species in the model. But methods that use presence only data also incorporate AUC as a common method to evaluate the model by incorporating likeness of pseudoabsene points spread around the study region. The AUC method of model evaluation is not dependent upon a preset threshold of the prevalence of positive sites vs the prevalence of absence sites in the dataset, but there are other methods of evaluation which are threshold dependent. One such method is called true skill statistic (TSS). The TSS compares the number of true positives against those that are ascribed to random guessing (i.e., false positives) (Allouche, Tsoar and Kadmon, 2006). One benefit of using TSS is that it is not affected by the numbers of prevalence of positives in the data set. The range of values for TSS vary from -1 to 1, where TSS values closer to 1 indicate that the model is predicting the actual number of positives correctly and not just due to random chance. Another method of evaluation that has been used traditionally is the Kappa statistic, which is quite like TSS, but used less frequently as it is very dependent upon the prevalence of positives in the data set (Fielding and Bell, 1997a; Allouche, Tsoar and Kadmon, 2006). Species Distribution modeling applied to vectors and pathogens Species distribution modeling has now become an important part of spatial epidemiology and is currently widely used to model pathogens and disease spread. In modern times the spread of different zoonotic pathogens has sparked the importance in making accurate predictions of future spread in pathogens. Species distribution modeling is important in epidemiology because it helps to visualize the current spatial patterns and trends in disease incidences and disease vectors or reservoirs; it helps understand which environmental factors support disease transmission and maintenance; and helps in predicting how risk of exposure may change in the future under different environmental and socioeconomic scenarios (Eisen and Eisen, 2011; Hay et al., 2013; Purse and 114 Golding, 2015). The eventual purpose of developing SDMs for pathogens and associated vectors is prevention of disease spread. In epidemiology SDMs have been used for multiple purposes including to plan national level intervention strategies, aid decision makers during assessments, help to inform individuals about decision to obtain vaccination/prophylaxis before travel (Eisen and Eisen, 2011; Hay et al., 2013). Understanding the ecology and epidemiology of the pathogen in a particular area or region is critical such that the process of developing and interpreting the SDM is better informed (Rogers and Randolph, 2003). Species distribution models applied to vector borne disease systems identify associations between the vector or the vector-borne disease data (could be presence/absence or abundance data) and environmental or socioeconomic predictor variables (Eisen and Eisen, 2011). The SDMs for vectors and their vector-borne diseases result in the development of maps that display potential suitable locations for the vector or the vector-borne disease to thrive, which can be extremely helpful for indicating areas of risk where surveillance data were lacking (Eisen and Eisen, 2011). These maps, which are still hypotheses, can enable public health officials and researchers to direct limited resources to monitor most efficiently areas of disease risk. These models can also be combined with human demographic data to assess the proportion of people that are potentially at a risk of exposure to vectors and vector-borne diseases (Eisen and Eisen, 2011). The earliest form of developing modern species distribution model to predict suitable habitats for disease-causing vectors was applied for mosquito control in New Orleans (National Aeronautics and Space Administration, 1973). The study was conducted by the National Aeronautics and Space Administration (NASA) and used color infrared photography to map vegetation to identify suitable habitats for mosquito Aedes sollicitans larval sites within marshes of New Orleans. Similar to the study by NASA, a study conducted in Michigan mapped forested 115 wetlands and open marshes to find suitable mosquito breeding habitats around the Saginaw Bay area of Michigan, which was then used directly for targeting mosquito control measures (Wagner et al., 1979). Several other studies have used early photointerpretation techniques to identify potential breeding habitats for various species of mosquitoes within Texas (Welch et al., 1989), Lousiana (Cibula, 1976; Barnes and Cibula, 1979), Nebraska, and South Dakota (Hayes et al., 1985). Building upon these early methods, land satellite data were used to model habitat features that are important in Rift Valley Fever activity in Kenya (Linthicum et al., 1987; Pope et al., 1992). After the introduction of West Nile Virus to the United States in the late 90’s, several studies were conducted to generate risk maps of West Nile virus (Brownstein et al., 2002) and predict suitable habitats in the US for several of the West Nile Virus mosquito vectors (Diuk-Wasser, Brown, et al., 2006; Larson et al., 2010; Rochlin et al., 2011). Many studies have been conducted using species distribution modeling to predict suitable habitats for the malaria mosquito vector Anopheles spp. as malaria remains to be one of the leading causes of death in the developing world (Guerra, Snow and Hay, 2006; Hay and Snow, 2006; Kulkarni, Desrochers and Kerr, 2010; Obsomer, Defourny and Coosemans, 2012; Gwitira et al., 2015, 2018; Akpan et al., 2018; Frak et al., 2020). Most of these studies have been conducted within the African continent as majority of malaria deaths are reported in within this region. The majority of these models use maximum entropy techniques to identify areas that have suitable habitat conditions for several species of Anopheles spp. (Kulkarni, Desrochers and Kerr, 2010; Obsomer, Defourny and Coosemans, 2012; Gwitira et al., 2015; Akpan et al., 2018). With the increase in invasive mosquito species such as Aedes aegypti and Aedes albopictus, which serve as vectors for viruses such as dengue, Japanese encephalitis, chikungunya and Zika, modeling potential suitable habitats have become vastly important. Many studies have been conducted globally to predict the distribution of these invasive 116 mosquitoes (e.g., Kobayashi, Nihei and Kurihaha, 2002; Brady et al., 2014; Kraemer et al., 2015; Messina et al., 2016; Kamal et al., 2018; Leta et al., 2018). Species distribution modeling of Ixodes scapularis Tick-borne diseases have become a global threat affecting humans, domestic animals, and livestock. In the United States alone in the past decade there has been a gradual increase in tickborne disease cases, where currently at least 75% of the vector borne disease cases reported to the Centers for Disease Control and Prevention (CDC) is attributed to tick borne pathogens (Rosenberg et al., 2018). Of those reported tick-borne disease cases, Lyme disease accounts for at least 82%, making Lyme disease the most common vector borne disease in the United States (Eisen and Eisen, 2018b; Rosenberg et al., 2018). Recently ~34,000 cases of Lyme disease are reported annually to the CDC (Rosenberg et al., 2018; N. C. for E. and Z. I. D. D. of V.-B. D. Centers for Disease Control and Prevention, 2022). Along with Lyme disease, other tickborne diseases such as anaplasmosis and ehrlichiosis have also increased in case numbers since 2011 (Rosenberg et al., 2018). The expansion of the geographic range of disease-vectoring ticks species, and the subsequent increase in tickborne pathogen detection awareness among clinicians and improved diagnostic methods have contributed to the increase in tickborne diseases cases in the US (Eisen and Eisen, 2018b; Sonenshine, 2018b) Of the tick species of interest in the United States, the eastern blacklegged tick (Ixodes scapularis) plays an important role in the spread of tickborne disease since it is the primary vector for several zoonotic pathogens including, six bacterial agents: Borrelia burgdorferi and B. mayonii, the causative agents of Lyme disease; Anaplasma phagocytophilum, the causative agent of human granulocytic anaplasmosis; Borrelia miyamotoi, the causative agent of hard tick relapsing fever borreliosis; and Ehrlichia muris eauclarensis, the causative agent of ehrlichiosis. Ixodes scapularis 117 is also the vector for a protozoan parasite, Babesia microti the causative agent of babesiosis, and a potentially fatal flavivirus, Powassan virus, the causative agent of Powassan virus encephalitis. The majority of I. scapularis borne disease cases are concentrated into two major foci in the United States, one being the upper Midwest and the other being the Northeast (Hahn et al., 2016). The earliest understanding of the range of I. scapularis show it to be predominantly in the southeastern US, from the Gulf Coast reaching north along the Atlantic Coast to southern Massachusetts, in the central Midwest to Iowa and Indiana, and few specimens from southern Ontario (Bishopp and Trembley, 1945). The earliest record for I. scapularis in the northern U.S. was in the 1920s near Cape Cod, Massachusetts (Spielman et al., 1985). Populations of I. scapularis were reported sporadically by the 1940s along the Northern Atlantic coast (Bishopp and Trembley, 1945; Eisen and Eisen, 2018b). After the 1970s more records of I. scapularis were reported along the New England coastline, in Rhode Island, southern New York, northwestern Wisconsin, and Ontario (Jackson and DeFoliart, 1970; Good, 1973; Watson and Anderson, 1976; Ruebush et al., 1977; Hyland et al., 2000). It is hypothesized that I. scapularis probably was distributed throughout the upper Midwest and Northeast pre-European settlement (Tsao et al. 2021). By the late 1800s and early 1900s, however, due to rapid deforestation and severe depopulation of the white-tailed deer (Odocoileus virginianus) (the main host for the reproductive stage), the abundance and distribution of I. scapularis was probably nearly extirpated in these areas except for a few refugia (Eisen et al., 2017; Spielman et al. 1985). But by the turn of the 20th century reforestation efforts began, along with managed hunting and reintroduction of the white-tailed deer, eventually resulting in the increase and spread of I. scapularis in the Northeast and upper Midwest U.S. (Spielman et al., 1985; Lane, Piesman and Burgdorfer, 1991; Dennis et al., 1998; Eisen et al., 2017; Eisen and Eisen, 2018b; Sonenshine, 2018b). While similar dynamics regarding forests and white-tailed deer 118 populations occurred in the southern U.S. (Paddock and Yabsley, 2007), effects on southern populations of I. scapularis have not been discussed in the literature (or our knowledge), perhaps in part because southern populations of blacklegged ticks historically have not been recognized as vectors of disease (Lane, Piesman and Burgdorfer, 1991) The increase in surveillance efforts have helped to map the distribution of I. scapularis within the eastern USA (Centers for Disease Control and Prevention (CDC), 2022). Dennis et al. 1998 published the first county-wide distribution map of I. scapularis within the US. This early distribution map was based on records collected through surveys and publicly available data, and it was at the county scale since most available data was by county, but also the scale at which public health metrics are reported state-wide and nationally (Dennis et al., 1998).This study was the first to introduce a scale to define establishment of ticks populations, where counties were considered to be “reported” if at least one tick of any life stage had been identified within one calendar year. Counties were considered to have “established” I. scapularis if at least 6 ticks from one life stage or 2 life stages (larvae, nymph, adult) had been detected within one calendar year (Dennis et al., 1998). This classification of defining tick populations is the current standard used by the CDC not just for I. scapularis, but for other vector tick species. Based on this classification method the distribution of I. scapularis has rapidly expanded with more counties reporting established populations (Eisen, Eisen and Beard, 2016; Beard, Eisen and Eisen, 2021). With this rapid expansion of I. scapularis populations, cases of Lyme disease and anaplasmosis cases have increased as well (Centers for Disease Control and Prevention, 2023b, 2023a). This increase in disease incidence is mainly due to the expansion of the I. scapularis population within the eastern USA. Interestingly, the cases of these diseases have also increased in western US., where Ixodes pacificus is the vector, whose known range has expanded beyond what was reported in Bishopp 119 and Trembley (1945), but not as dramatically as that of northern populations of I. scapularis (Dennis et al. 1998; Eisen et al. 2015). Thus, increased cases of these diseases may be due mainly to increased surveillance and reporting and local ecological changes. With this rapid expansion of I. scapularis, efforts to incorporate species distribution models to predict the spread of I. scapularis has become vital. Therefore, it is important to understand the importance of several different environmental factors that impact the survival and expansion of I. scapularis (Estrada-Peña, 2002). Global climatic changes likely will also influence the distribution of I. scapularis, and thus species distribution models may be able to help predict future disease risk. For example, climate change may create suitable environmental conditions allowing I. scapularis to invade into areas previously deemed unsuitable for I. scapularis survival; likewise, climate change may decrease the suitability of other areas, leading to range contraction. The suitable habitats in the United State for the development of I. scapularis seem to be undergoing a rapid geographic change allowing I. scapularis to survive in habitats where previously I. scapularis was not detected (Estrada-Peña, 2002). The earliest I. scapularis species distribution model was that of Guerra et al. (2002), which identified several environmental factors associated with suitable habitats for I. scapularis establishment primarily in Wisconsin and northern Illinois but also Menominee County in the Upper Peninsula of Michigan. Using a logistic regression model, Guerra et al. (2002) showed soil features such as bedrock geology, quaternary geology, soil order and texture; habitat features such as land cover, forest types and elevation; and climatic features such as annual precipitation and snow fall were important variables that determined habitat suitability for the introduction and establishment of I. scapularis. Although Wisconsin is recognized as the center of the major foci of I. scapularis, at the time the model was developed, the distribution of I. scapularis was still limited 120 to certain areas, and thus it was a very timely to use species distribution modeling to predict habitats in the rest of Wisconsin and northern Illinois (Guerra et al., 2002). Later, Foster and colleagues (Foster 2004) tested the ability of the model in Guerra et al. (2002) to predict suitable habitats for I. scapularis in Michigan, which led to the discovery of established populations of blacklegged tick in southwestern Michigan (Erik Scott Foster, 2004). Soon after the publication of Guerra et al. (2002); Brownstein et al (2003) published the first eastern U.S. species distribution map, which also was a logistic regression model but was based only on climate variables. Brownstein et al. (2003) found that climatic extremes and variation in humidity were major indicators of a suitable habitats. When looking at the expansion of I. scapularis, maximum temperatures, minimum temperatures and the vapor pressure played an important role determining the range expansion of I. scapularis within the United States (Brownstein, Holford and Fish, 2005). In addition to expansion into southern Canada, they predicted an expansion of I. scapularis by 2080 within areas of Virginia, North Carolina, Georgia, Minnesota, Iowa and Michigan (Brownstein, Holford and Fish, 2003). The largest species distribution map developed for I. scapularis was developed by Diuk- Wasser et al. (2006, 2010). The geographic extent of this species distribution model was from the 100th meridian to the Atlantic Ocean and from the border with Canada to the Gulf Coast. This area was divided into 2-degree grids, from which blacklegged ticks were systematically sampled by drag cloth over four years, where the zero-inflated negative binomial regression model was used. Because this map was focused on modeling the spatial risk of Lyme disease, it focused efforts on modeling the density of questing I. scapularis nymphs (Diuk-Wasser et al. 2010) and then the density of questing infected nymphs (Diuk-Wasser et al. 2012). Nymphs are the epidemiologically most important life stage. Interestingly, although I. scapularis is widespread through the southern 121 U.S., nymphs have a different questing behavior, such that they rarely contact humans, and thus Lyme disease risk is concentrated when northern populations of I. scapularis are. Thus, after the first year, after questing nymphs were rarely dragged in the southern areas, researchers focused and increased sampling in the northern areas above approximately the 39th parallel. They found that the most important factors predicting nymphal density were altitude, monthly mean vapor pressure deficit and spatial autocorrelation, but other factors such as forest fragmentation and soil texture were not. Since this study, there has not been another large-scale species distribution model developed estimating the risk of Lyme disease based on systematic active surveillance, perhaps because efforts to systematically sample for I. scapularis are too resource intensive. Species modeling of I. scapularis, however, has not stopped. As I. scapularis has been spreading, there have been more and more surveys for the tick conducted. Even though sampling efforts have not been conducted systematically, presence data are available and can be modeled. Two models were developed for the northern regions of I. scapularis’ range based on presence and absence of I. scapularis modeled at the county-scale (Hahn et al., 2016; Burtis et al., 2022). The courser scale does not allow for more accurate predictions about I. scapularis distribution at finer scales when tick densities and environmental factors may be heterogeneous throughout a county. But, for public health departments at the state and federal level, they provide some guidance for prevention measures for which the county level is probably adequate. These maps also provide state and local health departments guidance on how to allocate limited resources for surveillance. As more surveillance for ticks are conducted, more SDMs can be conducted at a finer scale either using presence absence-based models (Lippi, Gaff, White, St. John, et al., 2021; Kopsco et al., 2023) or using tick abundance or density of nymphs (Diuk-Wasser, Brown, et al., 2006; Diuk- Wasser et al., 2010). 122 Currently we see an increase in ticks and tickborne diseases (Bacon, Kugeler and Mead, 2008; Adams et al., 2017; Rosenberg et al., 2018). With the impact of climate change, habitat modification, habitat fragmentation, along with changes in hosts communities (Rocklöv and Dubrow, 2020; Couper, MacDonald and Mordecai, 2021; Nuttall, 2022), the spread of I. scapularis and other vector tick species such as Amblyomma americanum (lone star ticks), Dermacentor variabilis (the American dog tick) and A. maculatum (Gulf Coast tick) are predicted to change (Sonenshine 2018). As such, the risk for the pathogens they vector are also predicted to change. By developing SDMs for these tick species and even their pathogens if data allow, we can identify the distribution of potential suitable habitats for certain tick species, which will help guide future efforts to prevent ticks and tickborne diseases. Conclusion Michigan is a state that is at the leading edge of a I. scapularis invasion in the Upper Midwest (Dennis et al., 1998; Eisen, Eisen and Beard, 2016; Lantos et al., 2017; Burtis et al., 2022). Ixodes scapularis was first discovered in the 1990s in the Upper Peninsula in Menominee a county bordering Wisconsin (Strand, Walker and Merritt, 1992) and then in the early 2000s in the southwestern Lower Peninsula in Berrien a county at the edge of Michigan bordering Indiana (Erik Scott Foster, 2004). Since this initial discovery I. scapularis has been gradually spreading along the western coastline of Lake Michigan Northward, along the eastern coastline of Lake Huron and interior southern regions in the Lower Peninsula (Dennis et al., 1998; Hamer et al., 2010, 2014; Eisen, Eisen and Beard, 2016). Slower but continued spread has also occurred in the Upper Peninsula, especially in the western region. With increasing tick populations in Michigan over time we see that human case incidences have increased as well (Michigan Department of Health and Human Services, 2021, 2022). Because there is a gradual, persistent spread of I. scapularis, using 123 SDMs to predict the spread of I. scapularis will be important for surveillance and reducing public health risk. Our study in developing SDMs for I. scapularis based on the presence-absence data will guide future investigations into potential regions that have suitable sites for the establishment of I. scapularis which can then serve also to improve future models as well as guide prevention of tick-borne disease. 124 BIBLIOGRAPHY Adams, D.A. et al. (2015) ‘Summary of Notifiable Infectious Diseases and Conditions — United States, 2015’. Available at: https://www.cdc.gov/MMWR/ Akpan, G.E. et al. (2018) ‘Dominant malaria vector species in Nigeria: Modelling potential distribution of Anopheles gambiae sensu lato and its siblings with MaxEnt’, PLoS ONE, 13(10), pp. 1–15. Available at: https://doi.org/10.1371/journal.pone.0204233. Allouche, O., Tsoar, A. and Kadmon, R. (2006) ‘Assessing the accuracy of species distribution models: Prevalence, kappa and the true skill statistic (TSS)’, Journal of Applied Ecology, 43(6), pp. 1223–1232. Available at: https://doi.org/10.1111/j.1365-2664.2006.01214.x. Bacon, R.M., Kugeler, K.J. and Mead, P.S. (2008) ‘Surveillance for Lyme Disease-United States, 1992-2006.’, MMWR. CDC Surveillance summaries: Morbidity and mortality weekly report, pp. 1–9. Available at: https://www.cdc.gov/mmwr/preview/mmwrhtml/ss5710a1.htm (Accessed: 4 July 2023). Barnes, C.M. and Cibula, W.G. (1979) ‘Some implications of remote sensing technology in insect control programs including mosquitoes.’, Mosquito News, 39(2), pp. 271–282. Bishopp, F.C. and Trembley, H.L. (1945) ‘Distribution and Hosts of Certain North American Ticks’, The Journal of Parasitology, 31(1), p. 1. Available at: https://doi.org/10.2307/3273061. Brady, O.J. et al. (2014) ‘Global temperature constraints on Aedes aegypti and Ae. albopictus persistence and competence for dengue virus transmission’, Parasites & Vectors, 7(1), p. 338. Available at: https://doi.org/10.1186/1756-3305-7-338. Breiman, L. (2001) ‘Random forests’, Machine Learning, 45, pp. 5–32. Available at: https://doi.org/10.1007/978-3-030-62008-0_35. Brownstein, J.S. et al. (2002) ‘Spatial analysis of West Nile Virus: Rapid risk assessment of an introduced vector-borne zoonosis’, Vector Borne and Zoonotic Disease, 2(3). Available at: https://doi.org/https://doi.org/10.1089/15303660260613729. Brownstein, J.S., Holford, T.R. and Fish, D. (2003) ‘A climate-based model predicts the spatial distribution of the Lyme Disease vector Ixodes scapularis in the United States’, Environmental Health Perspectives, 111(9), pp. 1152–1157. Available at: https://doi.org/10.1289/ehp.6052. Burtis, J.C. et al. (2022) ‘Predicting distributions of blacklegged ticks (Ixodes scapularis), Lyme disease spirochetes (Borrelia burgdorferi sensu stricto) and human Lyme Disease cases in the eastern United States’, Ticks and Tick-borne Diseases, 13(5). Available at: https://doi.org/10.1016/j.ttbdis.2022.102000. Centers for Disease Control and Prevention (2023a) Epidemiology and Statistics: Anaplasmosis. Available at: https://www.cdc.gov/anaplasmosis/stats/index.html (Accessed: 4 July 2023). 125 Centers for Disease Control and Prevention (CDC) (2022) ‘Regions where ticks live’. Available at: https://www.cdc.gov/ticks/geographic_distribution.html (Accessed: 4 July 2023). Centers for Disease Control and Prevention (CDC). (2022) ‘Lyme Disease - reported cases by year, United States’. Available at: https://www.cdc.gov/lyme/datasurveillance/surveillance- data.html?CDC_AA_refVal=https%3A%2F%2Fwww.cdc.gov%2Flyme%2Fdatasurveillance%2F recent-surveillance-data.html (Accessed: 4 July 2023). Cibula, W.G. (1976) Application of remotely sensed multispectral data to automated analysis of marshland vegetation. Couper, L.I., MacDonald, A.J. and Mordecai, E.A. (2021) ‘Impact of prior and projected climate change on US Lyme Disease incidence’, Global Change Biology, 27(4), pp. 738–754. Available at: https://doi.org/10.1111/gcb.15435. Dennis, D.T. et al. (1998) ‘Reported Distribution of Ixodes scapularis and Ixodes pacificus (Acari: Ixodidae) in the United States’, Journal of Medical Entomology, 35(5), pp. 629–638. Available at: https://doi.org/10.1093/jmedent/35.5.629. Diuk-Wasser, M.A. et al. (2006) ‘Modeling the spatial distribution of mosquito vectors for West Nile Virus in Connecticut, USA’, Vector Borne and Zoonotic Disease, 6(3), pp. 115–126. Available at: https://doi.org/10.4324/9781003232490-9. Diuk-Wasser, M.A. et al. (2010) ‘Field and climate-based model for predicting the density of host-seeking nymphal Ixodes scapularis, an important vector of tick-borne disease agents in the eastern United States’, Global Ecology and Biogeography, 19(4), pp. 504–514. Available at: https://doi.org/10.1111/j.1466-8238.2010.00526.x. Drew, C.A., Wiersma, Y.F. and Huettmann, F. (2011) ‘Predictive species and habitat modeling in landscape ecology: Concepts and applications’, in Predictive Species and Habitat Modeling in Landscape Ecology: Concepts and Applications, pp. 1–313. Available at: https://doi.org/10.1007/978-1-4419-7390-0. Duarte, A., Whitlock, S.L. and Peterson, J.T. (2018) ‘Species distribution modeling’, Encyclopedia of Ecology, 6, pp. 189–198. Available at: https://doi.org/10.1016/B978-0-12- 409548-9.10572-X. Eisen, L. and Eisen, R.J. (2011) ‘Using geographic information systems and decision support systems for the prediction, prevention, and control of vector-borne diseases’, Annual Review of Entomology, 56, pp. 41–61. Available at: https://doi.org/10.1146/annurev-ento-120709-144847. Eisen, R.J. et al. (2017) ‘Tick-borne zoonoses in the United States: Persistent and emerging threats to human health’, ILAR Journal, 58(3), pp. 319–335. Available at: https://doi.org/10.1093/ilar/ilx005. Eisen, R.J. and Eisen, L. (2018) ‘The Blacklegged Tick, Ixodes scapularis: An Increasing Public 126 Health Concern’, Trends in Microbiology, 34(4), pp. 295–309. Available at: https://doi.org/10.1016/j.pt.2017.12.006.The. Eisen, R.J., Eisen, L. and Beard, C.B. (2016) ‘County-scale distribution of Ixodes scapularis and Ixodes pacificus (Acari: Ixodidae) in the continental United States’, Journal of Medical Entomology, 53(2), pp. 349–386. Available at: https://doi.org/10.1093/jme/tjv237. Elith, J. and Franklin, J. (2013) ‘Species Distribution Modeling’, Encyclopedia of Biodiversity: Second Edition, 6, pp. 692–705. Available at: https://doi.org/10.1016/B978-0-12-384719- 5.00318-X. Elith, J. and Leathwick, J.R. (2009) ‘Species distribution models: Ecological explanation and prediction across space and time’, Annual Review of Ecology, Evolution, and Systematics, 40, pp. 677–697. Available at: https://doi.org/10.1146/annurev.ecolsys.110308.120159. Elith, J., Leathwick, J.R. and Hastie, T. (2008) ‘A working guide to boosted regression trees’, Journal of Animal Ecology, 77(4), pp. 802–813. Available at: https://doi.org/10.1111/j.1365- 2656.2008.01390.x. Estrada-Peña, A. (2002) ‘Increasing habitat suitability in the United States for the tick that transmits Lyme Disease: A remote sensing approach’, Environmental Health Perspectives, 110(7), pp. 635–640. Available at: https://doi.org/10.1289/ehp.02110635. Fielding, A.H. and Bell, J.F. (1997) ‘A review of methods for the assessment of prediction errors in conservation presence/absence models’, Environmental Conservation, 24(1), pp. 38–49. Available at: https://doi.org/10.1017/S0376892997000088. Foster, E.S. (2004) ‘Ixodes scapularis (Acari: Ixodidae) and Borrelia burgdorferi in Southwest Michigan: Population ecology and verification of a geographic risk model’, Michigan State University. Frak, A.N. et al. (2020) ‘Leveraging big data for public health: Mapping malaria vector suitability in Malawi with Google Earth Engine’, PLoS ONE, 15(8 August), pp. 1–21. Available at: https://doi.org/10.1371/journal.pone.0235697. Good, N.E. (1973) ‘Ticks of Eastern Long Island: Notes on Host Relations and Seasonal Distribution’, Annals of the Entomological Society of America, 66(2), pp. 240–243. Available at: https://doi.org/10.1093/aesa/66.2.240. Grinnell, J. (1917) ‘The Niche-Relationships of the California Thrasher’, The Auk, 34(4), pp. 427– 433. Available at: https://doi.org/10.2307/4072271. Guerra, C.A., Snow, R.W., and Hay, S.I. (2006) ‘Defining the Global Spatial Limits of Malaria Transmission in 2005’, Advances in Parasitology, 62(05), pp. 157–179. Available at: https://doi.org/10.1016/S0065-308X(05)62005-2. 127 Guerra, M. et al. (2002) ‘Predicting the risk of Lyme Disease: Habitat suitability for Ixodes scapularis in the north central United States’, Emerging Infectious Diseases, 8(3), pp. 289–297. Available at: https://doi.org/10.3201/eid0803.010166. Gwitira, I. et al. (2015) ‘Modelled habitat suitability of a malaria causing vector (Anopheles arabiensis) relates well with human malaria incidences in Zimbabwe’, Applied Geography, 60, pp. 130–138. Available at: https://doi.org/10.1016/j.apgeog.2015.03.010. Gwitira, I. et al. (2018) ‘Application of GIS to predict malaria hotspots based on Anopheles arabiensis habitat suitability in Southern Africa’, International Journal of Applied Earth Observation and Geoinformation, 64(January 2017), pp. 12–21. Available at: https://doi.org/10.1016/j.jag.2017.08.009. Hahn, M.B. et al. (2016) ‘Modeling the geographic distribution of Ixodes scapularis and Ixodes pacificus (Acari: Ixodidae) in the contiguous United States’, Journal of Medical Entomology, 53(5), pp. 1176–1191. Available at: https://doi.org/10.1093/jme/tjw076. Hamer, S.A. et al. (2010) ‘Invasion of the Lyme Disease vector Ixodes scapularis: Implications for Borrelia burgdorferi endemicity’, EcoHealth, 7(1), pp. 47–63. Available at: https://doi.org/10.1007/s10393-010-0287-0. Hamer, S.A. et al. (2014) ‘Increased diversity of zoonotic pathogens and Borrelia burgdorferi strains in established versus incipient Ixodes scapularis populations across the Midwestern United States’, Infection, Genetics and Evolution, 27, pp. 531–542. Available at: https://doi.org/10.1016/j.meegid.2014.06.003. Hay, S.I. et al. (2013) ‘Global mapping of infectious disease’, Philosophical Transactions of the Royal Society B: Biological Sciences, 368(1614). Available at: https://doi.org/10.1098/rstb.2012.0250. Hay, S.I. and Snow, R.W. (2006) ‘The Malaria Atlas Project: Developing global maps of malaria risk’, PLoS Medicine, 3(12), pp. 2204–2208. Available at: https://doi.org/10.1371/journal.pmed.0030473. Hayes, R.O. et al. (1985) ‘Detection, identification, and classification of mosquito larval habitats using remote sensing scanners in earth-orbiting satellites.’, Bulletin of the World Health Organization, 63(2), pp. 361–74. Available at: http://www.ncbi.nlm.nih.gov/pubmed/2861917. Hyland, K.E. et al. (2000) ‘Records of ticks (acari: Ixodidae) parasitizing birds (Aves) in Rhode Island, USA’, International Journal of Acarology, 26(2), pp. 183–192. Available at: https://doi.org/10.1080/01647950008684185. Jackson, J.O. and DeFoliart, G.R. (1970) ‘Ixodes scapularis Say in northern Wisconsin.’, Journal of medical entomology, 7(1), pp. 124–125. Available at: https://doi.org/10.1093/jmedent/7.1.124. Jiménez-Valverde, A. (2012) ‘Insights into the area under the receiver operating characteristic 128 curve (AUC) as a discrimination measure in species distribution modelling’, Global Ecology and Biogeography, 21(4), pp. 498–507. Available at: https://doi.org/10.1111/j.1466- 8238.2011.00683.x. Kamal, M. et al. (2018) ‘Mapping the global potential distributions of two arboviral vectors Aedes aegypti and Ae. Albopictus under changing climate’, PLoS ONE, 13(12), pp. 1–21. Available at: https://doi.org/10.1371/journal.pone.0210122. Kobayashi, M., Nihei, N. and Kurihaha, T. (2002) ‘Analysis of northern distribution of Aedes albopictus (Diptera: Culicidae) in Japan by geographical information system’, Journal of Medical Entomology, 39(1), pp. 4–11. Available at: https://doi.org/10.1603/0022-2585-39.1.4. Kopsco, H.L. et al. (2023) ‘Current and Future Habitat Suitability Models for Four Ticks of Medical Concern in Illinois, USA’, Insects, 14(3), p. 213. Available at: https://doi.org/10.3390/insects14030213. Kraemer, M.U.G. et al. (2015) ‘The global distribution of the arbovirus vectors Aedes aegypti and Ae. Albopictus’, eLife, 4(JUNE2015), pp. 1–18. Available at: https://doi.org/10.7554/eLife.08347. Kulkarni, M.A., Desrochers, R.E. and Kerr, J.T. (2010) ‘High resolution niche models of malaria vectors in Northern Tanzania: A new capacity to predict malaria risk?’, PLoS ONE, 5(2). Available at: https://doi.org/10.1371/journal.pone.0009396. Lane, R.S., Piesman, J. and Burgdorfer, W. (1991) ‘Lyme Borreliosis: Relation of Its Causative Agent to Its Vectors and Hosts in North America and Europe’, Annual Review of Entomology, 36(1), pp. 587–609. Available at: https://doi.org/10.1146/annurev.en.36.010191.003103. Lantos, P.M. et al. (2017) ‘Geographic expansion of Lyme disease in Michigan, 2000-2014’, Open Forum Infectious Diseases, 4(1), pp. 1–5. Available at: https://doi.org/10.1093/oid/ofw269. Larson, S.R. et al. (2010) ‘Ecological niche modeling of potential West Nile Virus vector mosquito species in Iowa’, Journal of Insect Science, 10(110), pp. 1–17. Available at: https://doi.org/10.1673/031.010.11001. Leta, S. et al. (2018) ‘Global risk mapping for major diseases transmitted by Aedes aegypti and Aedes albopictus’, International Journal of Infectious Diseases, 67, pp. 25–35. Available at: https://doi.org/10.1016/j.ijid.2017.11.026. Li, X. and Wang, Y. (2013) ‘Applying various algorithms for species distribution modelling’, Integrative Zoology, 8(2), pp. 124–135. Available at: https://doi.org/10.1111/1749-4877.12000. Linthicum, K.J. et al. (1987) ‘Detection of Rift Valley Fever viral activity in Kenya by satellite remote sensing imagery’, Science, 235(4796), pp. 1656–1659. Available at: https://doi.org/10.1126/science.3823909. Lippi, C.A. et al. (2021) ‘Exploring the Niche of Rickettsia montanensis (Rickettsiales: 129 Rickettsiaceae) Infection of the American Dog Tick (Acari: Ixodidae), Using Multiple Species Distribution Model Approaches’, Journal of Medical Entomology, 58(3), pp. 1083–1092. Available at: https://doi.org/10.1093/jme/tjaa263. Messina, J.P. et al. (2016) ‘Mapping global environmental suitability for Zika virus’, eLife, 5(APRIL2016), pp. 1–19. Available at: https://doi.org/10.7554/eLife.15272. Michigan Department of Health and Human Services (2021) ‘Michigan Trends in Tickborne Disease, 2016-2020’. Available at: https://www.michigan.gov/- /media/Project/Websites/emergingdiseases/Folder3/2021_Tickborne_Disease_Summary_Report.p df?rev=a77a79a5ca16467ebeef4a41c9272e55 (Accessed: 4 July 2023). Michigan Department of Health and Human Services (2022) ‘Michigan Emerging and Zoonotic Disease Surveillance Summary 2021’. Available at: https://www.michigan.gov/- /media/Project/Websites/emergingdiseases/EZID_Annual_Surveillance_Summary.pdf?rev=c41c1 aa053754235bc29f3d86c24b0c8 (Accessed: 4 July 2023). National Aeronautics and Space Administration (1973) ‘The Use of Remote Sensing in Mosquito Control. Nuttall, P.A. (2022) ‘Climate change impacts on ticks and tick-borne infections’, Biologia, 77(6), pp. 1503–1512. Available at: https://doi.org/10.1007/s11756-021-00927-2. Obsomer, V., Defourny, P. and Coosemans, M. (2012) ‘Predicted Distribution of Major Malaria Vectors Belonging to the Anopheles dirus Complex in Asia: Ecological Niche and Environmental Influences’, PLoS ONE, 7(11). Available at: https://doi.org/10.1371/journal.pone.0050475. Paddock, C.J. and Yabsley, M.J. (2007) ‘Ecological havoc, the rise of the white-tailed deer, and the emergence of Amblyomma americanum-associated zoonoses in the United States’, in J.A. Richt (ed.) The biology, circumstances and the consequences of cross-species transmission, pp. 289–324. Phillips, S.J., Anderson, R.P. and Schapire, R.E. (2006) ‘Maximum entropy modeling of species geographic distributions’, Ecological Modelling, 190, pp. 231–259. Available at: https://doi.org/10.1016/j.ecolmodel.2005.03.026. Pope, K.O. et al. (1992) ‘Identification of central Kenyan Rift Valley Fever virus vector habitats with landsat TM and evaluation of their flooding status with airborne imaging radar’, Remote Sensing of Environment, 40(3), pp. 185–196. Available at: https://doi.org/10.1016/0034- 4257(92)90002-2. Purse, B. V. and Golding, N. (2015) ‘Tracking the distribution and impacts of diseases with biological records and distribution modelling’, Biological Journal of the Linnean Society, 115(3), pp. 664–677. Available at: https://doi.org/10.1111/bij.12567. Rochlin, I. et al. (2011) ‘Predictive mapping of human risk for West Nile Virus (WNV) based on 130 environmental and socioeconomic factors’, PLoS ONE, 6(8). Available at: https://doi.org/10.1371/journal.pone.0023280. Rocklöv, J. and Dubrow, R. (2020) ‘Climate change: an enduring challenge for vector-borne disease prevention and control’, Nature Immunology. Nature Research, pp. 479–483. Available at: https://doi.org/10.1038/s41590-020-0648-y. Rogers, D.J. and Randolph, S.E. (2003) ‘Studying the global distribution of infectious diseases using GIS and RS’, Nature Reviews Microbiology, 1(3), pp. 231–237. Available at: https://doi.org/10.1038/nrmicro776. Rosenberg, R. et al. (2018) ‘Vital Signs: Trends in Reported Vectorborne Disease Cases — United States and Territories, 2004–2016’, MMWR. Morbidity and Mortality Weekly Report, 67(17), pp. 496–501. Available at: https://doi.org/10.15585/mmwr.mm6717e1. Ruebush, T.K. et al. (1977) ‘Human Babesiosis on Nantucket Island’, New England Journal of Medicine, 297(15), pp. 825–827. Available at: https://doi.org/10.1056/nejm197710132971511. Sonenshine, D.E. (2018) ‘Range expansion of tick disease vectors in North America: Implications for spread of tick-borne disease’, International Journal of Environmental Research and Public Health, 15(3), pp. 1–9. Available at: https://doi.org/10.3390/ijerph15030478. Spielman, A. et al. (1985) ‘Ecology of Ixodes dammini-borne human babesiosis and Lyme Disease’, Annual Review of Entomology, 13(1), pp. 439–460. Available at: https://doi.org/10.1146/annurev.en.30.010185.002255. Stockwell, D. and Peters, D. (1999) ‘The GARP modelling system: Problems and solutions to automated spatial prediction’, International Journal of Geographical Information Science, 13(2), pp. 143–158. Available at: https://doi.org/10.1080/136588199241391. Strand, M.R., Walker, E.D. and Merritt, R.W. (1992) ‘Field studies on Ixodes dammini in the Upper Peninsula of Michigan’, Vector Control Bulletin of North Central States, 1, pp. 11–18. Wagner, V.E. et al. (1979) ‘Remote sensing: a rapid and accurate method of data acquisition for a newly formed mosquito control district.’, Mosquito News, 39(2), pp. 283–287. Watson, T.G. and Anderson, R.C. (1976) ‘Ixodes scapularis Say on white-tailed deer (Odocoileus virginianus) from Long Point, Ontario.’, Journal of Wildlife Diseases, 12(1), pp. 66–71. Available at: https://doi.org/10.7589/0090-3558-12.1.66. Welch, J.B. et al. (1989) ‘Use of aerial color infrared photography as a survey technique for Psorophora columbiae oviposition habitats in Texas rice lands.’, Journal of the American Mosquito Control Association, 5(2), pp. 147–60. Available at: http://www.ncbi.nlm.nih.gov/pubmed/2568391. Whittaker, R.H., Levin, S.A. and Root, R.B. (1973) ‘Niche, Habitat , and Ecotope’, The American Naturalist, 107(955), pp. 321–338. Available at: https://www.jstor.org/stable/2459534 (Accessed: 131 4 July 2023). Zhang, J. and Li, S. (2018) ‘A Review of Machine Learning Based Species’ Distribution Modelling’, Proceedings - 2017 International Conference on Industrial Informatics - Computing Technology, Intelligent Technology, Industrial Information Integration, ICIICII 2017, pp. 199– 206. Available at: https://doi.org/10.1109/ICIICII.2017.76. 132 CHAPTER 5: SPECIES DISTRIBUTION MODELING OF THE BLACKLEGGED TICKS (IXODES SCAPULARIS) IN MICHIGAN ABSTRACT The increase in Ixodes scapularis-borne diseases have been attributed to the rapid expansion of the tick throughout the Northeast and the Upper Midwest. In the Upper Midwest, historically endemic populations of I. scapularis were first detected in Wisconsin in 1965, after which populations spread throughout the state. In comparison, in the neighboring state of Michigan, I. scapularis was first discovered in the early 1990s in the Upper Peninsula, in a county bordering Wisconsin, and in the early 2000s in the southwestern corner of the Lower Peninsula. As in Wisconsin, populations of I. scapularis have been steadily expanding in Michigan. With this invasion occurring in real time, we examined 1) the spread of I. scapularis; 2) predicted the environmental predictors important in those suitable habitats by developing the models, and 3) predicted the distribution of suitable habitats for I. scapularis in Michigan. We sampled for ticks using drag sampling from 2017 to 2021, where we sampled in 315 sites, although sampling sites differed among years. Due to the Lower and Upper Peninsula having very different ecological, climatic, and geological features we modelled each Peninsula separately using two modeling methods, a regression-based approach using a generalized linear model (GLM) and a machine learning method called maximum entropy (MaxEnt) modeling. We began with 29 initial environmental predictors, which was reduced to 8 for the Lower Peninsula and 9 for the Upper Peninsula in the final GLM models. The final MaxEnt models included 20 predictors for the Lower Peninsula and 18 for the Upper Peninsula. For the climatic predictors, variables related to temperature, humidity and for the ecological predictors features related to soil, elevation and forest types were important. 133 In evaluating each model, the area under the curve values (AUC) for the GLM and MaxEnt models in the Lower Peninsula were 0.78 and 0.94 respectively, while for the Upper Peninsula they were 0.85 and 0.93 respectively, indicating that all models performed better at classifying sites as positive (or negative) for I. scapularis than by random chance. The true skill statistics (TSS) values for the Lower Peninsula GLM and MaxEnt were 0.41 and 0.45, while for the Upper Peninsula they were 0.64 and 0.35 respectively, again, for all models indicating the presence sites were actual presence sites. In both modeling methods similar regions were identified to have high to low suitability habitats for I. scapularis. Through this study we hoped to find suitable regions in the Upper and Lower Peninsula where we can guide our surveillance effort to detect established populations of I. scapularis. With the increase in Lyme disease cases in Michigan it is important to predict where suitable sites for establishment of I. scapularis are found in Michigan to guide prevention and control strategies as soon as possible. Because Michigan is at the leading edge of an invasion of I. scapularis, model predictions may change, where some of the habitats currently predicted as unsuitable may become suitable for I. scapularis. When training the models, the data we used may have sites that are “false negatives” where I. scapularis was not detected because the tick still has not gotten a chance to get to those sites, to invade and become established yet. Surveillance and/or experimental tick survivorship studies conducted in the future in these putatively unsuitable regions can test these hypotheses. Keywords: Ixodes scapularis, species distribution modeling, Michigan, environmental predictors 134 INTRODUCTION Spread of I. scapularis in the Great Lakes region The blacklegged tick, Ixodes scapularis, is the vector for several important tick-borne pathogens including the Borrelia burgdorferi, B. miyamotoi and Anaplasma phagocytophilum, Babesia microti, and Powassan virus (Nelder et al., 2016; Eisen and Eisen, 2018a; Wolf, Watkins and Schwan, 2020). The historic range of I. scapularis encompasses much of the eastern United States, although populations have been expanding into new areas, particularly along the northern edge of their range (Eisen and Eisen, 2018a). The abundance of I. scapularis has also been increasing through much of the northern range (Dennis et al., 1998; Eisen, Eisen and Beard, 2016; Eisen and Eisen, 2018a; Fleshman et al., 2021). Ixodes scapularis has been considered endemic to parts of the Midwest and the Northeast for more than half a century (Dennis et al., 1998). It has been hypothesized that land use changes (Barbour and Fish, 1993; Pfäffle et al., 2013; VanAcker et al., 2019; Diuk-Wasser, Vanacker and Fernandez, 2021), habitat modifications (Pfäffle et al., 2013; Couper et al., 2020; Swei et al., 2020), increased deer population densities (Witmer and Decalesta, 1991; Khatchikian et al., 2015; Kugeler et al., 2016; Gourley et al., 2018; Fish, 2021), and possibly climate change (Süss et al., 2008; Ostfeld and Brunner, 2015; Ogden and Lindsay, 2016; Bouchard et al., 2019; Rocklöv and Dubrow, 2020; Nuttall, 2022) allowed ticks from limited populations to expand throughout these regions. Both the geographic expansion of tick populations and increased abundance have likely contributed to the rising incidence of human tick-borne diseases throughout the eastern United States (Adams et al., 2017; Eisen and Eisen, 2018a). The first populations of I. scapularis recorded from the Great Lakes region were from northcentral Wisconsin in the early 1960s (Jackson and DeFoliart, 1970; Gardner et al., 2020). 135 Since then, populations of I. scapularis have been spreading throughout the state and into neighboring contiguous states to the west and south (Bouseman et al., 1990; Hamer et al., 2010; Khatchikian et al., 2015; Fish, 2021). In the neighboring state to the east, Michigan, the first established population of I. scapularis was discovered in the Upper Peninsula in a county bordering Wisconsin (Menominee County) in the late 1980’s (Strand, Walker and Merritt, 1992). In the early 2000’s, a population of I. scapularis was found in the southwest corner of the Lower Peninsula (Erik Scott Foster, 2004). Between 2004 to 2008, populations of I. scapularis expanded in Michigan at a faster rate northward through counties adjacent to Lake Michigan when compared to the rate of inland spread (Hamer et al., 2010). But by 2016, established populations of I. scapularis were detected further inland in the southern regions of the Lower Peninsula and were later found at sites in and near the “Thumb” region adjacent to Lake Huron (Eisen, Eisen and Beard, 2016; Lantos et al., 2017; Fleshman et al., 2021). Since the first population of I. scapularis was discovered in the Upper Peninsula of Michigan in 1990, populations of I. scapularis were detected in both the westernmost and easternmost counties of Michigan by 1998 (Dennis et al., 1998). By 2016, populations had spread even further into eastern areas of the Upper Peninsula throughout many counties (Eisen, Eisen and Beard, 2016). Currently, I. scapularis populations have been reported throughout the Upper Peninsula (Fleshman et al., 2021). Species Distribution Models and Ecological Associations for I. scapularis With the constant spread of I. scapularis populations into new geographic regions, and the subsequent spread or increase of associated diseases, there is a need to identify potential habitats along the expansion front of I. scapularis. Therefore, species distribution models (also referred to as habitat suitability models) have become an increasingly useful tool to identify climatic and ecological factors that may be associated with tick occurrence (Guerra et al., 2002; Bunnell et al., 136 2003; Lubelczyk et al., 2004; Diuk-Wasser et al., 2010; Johnson et al., 2016) Species distribution models (SDMs) can be used to predict additional areas into which the tick may eventually become established and be of public health concern (Illoldi-Rangel et al., 2012; Feria-Arroyo et al., 2014; Gabriele-Rivet et al., 2015; Lieske and Lloyd, 2018; Soucy et al., 2018; Kessler, Ganser and Glass, 2019; Slatculescu et al., 2020; Glass, Ganser and Kessler, 2021; Kopsco et al., 2023). Species distribution models therefore can be useful in identifying potential suitable habitats in regions surveillance has either been not carried out or have been under sampled especially along the I. scapularis expansion front. All the described usages of SDMs will ultimately be used to inform public health considerations. Identifying climatic and ecological correlates to I. scapularis occurrence is a critical first step in identifying potential habitat in emergent areas, such as Michigan. Species distribution models will help in identifying potential climatic and ecological correlates that are important for I. scapularis occurrence or establishments (i.e., two life stages of I. scapularis or greater than 5 individual I. scapularis of any life stage detected within one calendar year) especially along the expansion front of I. scapularis. Climatic and ecological factors likely play an important role in the establishment and spread of I. scapularis (Estrada-Pea, 2001; Estrada-Peña, 2002; Ostfeld et al., 2006; Diuk-Wasser et al., 2010; Johnson et al., 2016; Ginsberg et al., 2017; Gardner et al., 2020). Ticks are sensitive to humidity and are prone to desiccation at low humidity (Vail and Smith, 1998; Berger et al., 2014; Elias et al., 2021). Ticks are also highly sensitive to variation in temperature. At high temperatures ticks are susceptible to desiccation, there is a reduction in oviposition success in ticks, and high temperatures will decrease ticks host-seeking activities (Needham and Teel, 1991; Duffy and Campbell, 1994; Eisen et al., 2016) and at extreme low 137 temperatures there may be a delay in development rates, and increase mortality of I. scapularis(Lindsay et al., 1995; Vandyk et al., 1996; Eisen et al., 2016). Precipitation pattern changes are also suggested to affect ticks densities and occurrence. At typical precipitation for example rain, the conditions are kept moist which will promote tick survival (Burtis et al., 2016). Precipitation in the winter seasons which is typically snow, greater snowfall will increase snowpack densities keeping the ground level conditions moist and warm and tick will be sheltered from desiccation and the cold (Linske et al., 2019; Volk et al., 2022). With climate change these conditions seem to demonstrate frequent fluctuations which may affect the ability of ticks to survive. Local ecological characteristics are also likely to play a role in enabling ticks to survive. Many studies have shown that I. scapularis are found within deciduous woods with abundant shrubs and relatively thick layers of leaf litter (Ginsberg and Zhioua, 1996; Ginsberg et al., 2004; Linske et al., 2019). Understory and leaf litter cover likely help maintain moist and cool microhabitats needed by I. scapularis to avoid desiccation (Ginsberg and Ewing, 1989; Adler et al., 1992; Ostfeld et al., 1995; Lubelczyk et al., 2004). Despite coniferous cover not being the most optimal habitat type, I. scapularis have been found in habitats with coniferous cover, likely because of their wide niche range (Lindsay et al., 1999; Lubelczyk et al., 2004; Elias et al., 2006, 2022; Coyle et al., 2013; Lee et al., 2014). Coniferous forests may be less optimal when compared to deciduous forests, however, since they often provide minimal leaf litter, are associated with drier conditions (Ostfeld et al., 1995; Lindsay et al., 1998; Schulze, Jordan and Hung, 1998; Guerra et al., 2002). Another important ecological feature that may affect a tick’s ability to survive, is soil composition. Models have predicted that the presence of sandy soil, which allows drainage of excess water, increases I. scapularis suitability and survivorship because it prevents the 138 introduction of fungi or parasitic nematodes on I. scapularis as they overwinter (Kitron et al., 1992; Morgan et al., 1994; Bertrand and Wilson, 1996; Lindsay et al., 1998; Schulze, Jordan and Hung, 1998; Guerra et al., 2002). Acidic soils with high clay content are predicted to be unsuitable since they do not allow drainage of excess water (Guerra et al., 2002). Using SDMs to predict suitable habitat in areas of recent expansion Since Michigan is a state with an ongoing invasion of I. scapularis (Dennis et al., 1998; Eisen, Eisen and Beard, 2016; Lantos et al., 2017; Fleshman et al., 2021), developing habitat suitability models would be beneficial because it will enable us to predict suitable habitats along the expansion front which will help in planning our surveillance efforts to target regions with suitable habitats. Habitat suitability models will also help in informing the public and healthcare workers of the future risk in ticks and tickborne diseases. The first species distribution model for the Great Lakes region was developed using occurrence records for I. scapularis in Wisconsin from 1996 to 1998 (Guerra et al., 2002). According to Guerra et al. (2002), upland forests composed predominantly of oak with sandy soils were high suitability habitats for I. scapularis. When the Guerra et al. 2002 model was projected onto Michigan, areas that were predicted to contain highly suitable habitat predominantly distributed in the northern Lower Peninsula and several areas in the southwest along the Lake Michigan shoreline and inland areas along the “Thumb” region of Michigan (Erik Scott Foster, 2004). Using species distribution models developed from endemic areas and projecting those model predictions onto region of recent emergence is one possible approach that may help to identify suitable habitat patches along the expansion front. Extrapolating data trends from endemic zones, however, may not capture differences in habitat and climatic conditions associated with novel and distinct geographic areas. Michigan, for example, has eco-physiographic contexts that 139 diverge in some respects from neighboring states. While much of northern Wisconsin and the Upper Peninsula of Michigan fall into the Northern Lakes and Forests ecotype (Omernik and Griffith, 2014), southern areas deviate. Southwestern Wisconsin, for example, falls into the Driftless Area ecoregion, where glacial processes had less pronounced effects on the surrounding landscape, whereas southwestern Michigan is predominantly characterized as Michigan/Indiana Drift Plains (Appendix, Figure 5.9), with predominant landforms and forest types more closely associated with post-glacial drift processes (Leverett and Taylor, 1915; Bertrand and Wilson, 1996; Clayton, Attig and Mickelson, 2001; Fisher, Jol and Boudreau, 2005; Omernik and Griffith, 2014; Dickmann and Leefers, 2016). Another major difference is in central Wisconsin there is a large region of northcentral hardwood forests where in Michigan this ecoregion is restricted to the Traverse City Bay area on northwestern region bordering Lake Michigan. So, while many ecological similarities exist between Michigan and other states in the Great Lakes region, models developed in the endemic areas may not necessarily capture differences in ecological context and which may have important effects on the distributional pattern of I. scapularis in the state. Michigan is also geographically distinct from other states in the Upper Midwest because it is comprised of two geographically distinct peninsulas, which are both surrounded by The Great Lakes. This peninsular geography may create heterogeneity in climatic and ecological conditions relevant to the survivorship and distribution of I. scapularis throughout the state. Areas more proximate to the Great lakes have more stable temperatures, precipitation patterns, and snow fall throughout the winter season relative to inland areas (Scott and Huff, 1996). The southern region of the Lower Peninsula is also ecologically distinct from the Upper Peninsula (Omernik and Griffith, 2014). Much of the southern regions of the Lower Peninsula belongs to either the Southern Michigan/Northern Indiana drift plains (Omernik and Griffith, 2014) in the west and 140 Huron/Erie Lake plains on the east, while the northern regions resemble the Upper Peninsula which predominantly belong to the Northern Lakes and Forests ecoregion (Omernik and Griffith, 2014). The southern region of the Lower Peninsula is dominated by deciduous hardwood tree species like oak, beech, maple, and hickory, while the northern regions of the Lower Peninsula dominated by pure conifers, mixed conifers, and hardwood forests (Dickmann and Leefers, 2016). In the Upper Peninsula the eastern region is much like the northern regions of the Lower Peninsula, but with much more wetlands, which are dominated by swamp conifer forests, while the western region of the Upper Peninsula is dominated by a mix of hardwoods and conifers (Dickmann and Leefers, 2016). Human densities are also higher in the Lower Peninsula (U.S. Census Bureau, 2022), leading to greater degrees of forest fragmentation, urbanization, and agricultural land use (Dickmann and Leefers, 2016). Lastly, and most importantly this unique heterogenous geography and ecology of the Lower and Upper Peninsulas has likely affected the expansion dynamics of I. scapularis, such that the populations found in each peninsula represents a distinct focus where populations emerged from different regions. The objective of this study was to identify climatic and habitat conditions associated with I. scapularis occurrence in the Lower and Upper Peninsulas of Michigan and to predict suitable habitat patches. We used two different methods to model the distributions of I. scapularis in Michigan, which each use a slightly different dataset and a different set of assumptions. We used Michigan statewide surveillance data collected from 2017 to 2021 to identify ecological correlates to I. scapularis occurrence and model suitable habitat. We fit models for the Lower and Upper Peninsulas separately due to differences in ecology, geography, and invasion dynamics between the two peninsulas. We expect these will improve the understanding of where suitable sites for I. 141 scapularis are found and what environmental conditions are important in facilitating the spread and establishment of populations throughout Michigan. MATERIALS AND METHODS Sampling sites and tick occurrence data Sampling efforts were primarily focused to areas with upland deciduous forests, because of previously documented associations between I scapularis and deciduous forest cover (Ostfeld et al., 1995; Lubelczyk et al., 2004; Randolph, 2004; Pfäffle et al., 2013; Linske et al., 2019). We also sampled in mixed hardwood forests, which comprised of deciduous hardwood trees mixed with coniferous trees, and a few sites were mainly coniferous. Our sites (total sites = 315; Figure 5.1) consisted of public lands ranging from national and state forests, state, county, and city parks, nature conservancies, nature centers, university properties, and a few individually owned properties (e.g., private residences). Questing ticks were sampled from April to November, 2017 to 2021, targeting the peak activity periods of nymphal and adult I. scapularis. Ticks were sampled by drag cloth (Hamer et al. 2010), where a 1 m2 white-colored corduroy or flannel cloth or 0.75 m2 canvas cloth with “fingers” weighed down with curtain weights was dragged on the ground over the leaf litter (Rulison et al., 2013; Centers for Disease Control and Prevention, 2019). At most sites, samples were collected along 1600 m transects that were located along the margins of hiking trails because trail sampling best represents risk of tick-borne exposure to the public. In some cases, sampling 1600 m was not possible, but all sites exceeded the 750 m minimum drag distance suggested by the Centers for Disease Control tick surveillance recommendations (Centers for Disease Control and Prevention, 2019). Field collected ticks were stored in 90% ethanol, and species identification was confirmed in at the Michigan State University laboratory using dichotomous keys (Clifford, 142 Anastos and Elbl, 1961; Sonenshine, 1979; Keirans and Litwak, 1989; Durden and Keirans, 1997). Surveillance records were used to develop site- and county-level distribution maps using ArcGIS Pro 3.1 (ESRI, Redlands, CA, USA). Species distribution models: Environmental predictors We selected 29 environmental predictors (Table 5.1) for our initial model matrix, all of which were commonly evaluated in previous habitat association studies for I. scapularis (Springer et al., 2015; Hahn et al., 2016; Lippi, Gaff, White and Ryan, 2021; Lippi, Gaff, White, St. John, et al., 2021; Bacon et al., 2022; Kopsco, Smith and Halsey, 2022). All the environmental predictors were gridded geospatial data stored as rasters. Since the source and original resolution varied among datasets, all rasters were reprojected using a USA Albers’s Equal Area Conic projection (USGS version), resampled to a spatial resolution of 1 km2, and finally masked and cropped to the extent of Michigan using the raster package (version 3.6-20), in RStudio (version 2023.03.1). Climatic predictors included the 19 bioclimatic variables obtained from WorldClim version 2.1 (Fick and Hijmans, 2017). The bioclimatic variables are a set of variables based on temperature and precipitation measurements taken from global weather stations and averaged across the years from 1970 to 2000, which are meant to represent annual mean global trends, along with seasonality, extreme environmental factors and limiting environmental factors (Hijmans et al., 2005; Fick and Hijmans, 2017). Michigan is at the northern extent of the distribution of I. scapularis, making it more likely that these populations will experience greater seasonal variation in temperature and greater amounts of snowfall relative to southern endemic areas. Because of this, we included two additional climatic predictors, snow water equivalent and cumulative growing degree days, to represent climatic factors that may affect Michigan’s tick populations. Snow water equivalent 143 represents the accumulation of snow during winter months. Snow accumulation insulates the ground, maintaining more humid conditions and higher temperatures than exposed ground (Decker et al., 2003; Templer et al., 2012). This insulating effect likely helps to increase survivorship of I. scapularis in winter months (Linske et al., 2019; Volk et al., 2022). Cumulative growing degree days is another important measure that affects the development of I. scapularis where temperature changes affect the molting success and molting times of I. scapularis (Brunner et al., 2023). We calculated cumulative growing degree days for Michigan by taking the mid-range temperature and then secondarily evaluating the difference between the annual mid-range temperature and a minimum temperature threshold of 0 ºC (McMaster and Wilhelm, 1997). Snow water equivalent and maximum and minimum temperatures were obtained from the Daymet database (version 4; provided courtesy of Oak Ridge National Laboratory, Oak Ridge, TN, USA; Thornton et al., 2022, 1997). In addition to climate predictor rasters, we also included three categories of ecological predictors in our model: (1) landcover and habitat, (2) soil properties, and (3) host presence. Landcover and habitat predictors included the National Landcover Database (NLCD) land cover types and forest groups. The NLCD land cover type is a raster dataset where each pixel indicates the predominant cover type of an area, with 16 options including open area or water, degree of urban development, barren land, forest cover, shrubland, herbaceous cover and grassland, croplands, and wetlands (Yang et al., 2018; Dewitz and U.S. Geological Survey, 2021; Wickham et al., 2021) provided courtesy of the Multi-Resolution Land Characteristics Consortium (Sioux Falls, SD, USA). The subcategories (16 layers) were combined to form seven raster layers representing the major cover types listed above. In addition to major land-use categories, we also included major forest group classifications in our final model set. The forest groups data used 144 advanced resolution radiometer imagery to represent 145 forest types spread across the United States (Ruefenacht et al., 2008) provided courtesy of USDA Forest Service Geodata Clearinghouse (United States Department of Agriculture, Washington D.C., USA). The forest groups were separated to form six raster layers: (1) groups comprising of white, red and jack pine forest groups, (2) spruce and fir forest groups, (3) oak and hickory forest groups, (4) elm, ash, and cottonwood forest groups, (5) maple, beech, and birch forest groups, and (6) aspen and birch forest groups. We incorporated information on the following soil properties in the model set: (1) soil organic carbon density at a depth of 0 to 5 cm, (2) soil water capacity until wilting point at a depth of 0 cm, (3) soil sand content, and (4) soil clay content, which were downloaded as gridded geospatial layers from SoilGrids (Poggio et al., 2021) provided courtesy of the International Soil Reference Information Centre (ISRIC) (Wageningen, The Netherlands). The soil organic carbon density at a depth of 0 to 5cm represents the carbon content of the topsoil layer, mainly consisting of leaf litter, which is likely to be beneficial for the survival of I. scapularis (Linske et al., 2019; Volk et al., 2022). The soil water capacity until wilting point represents the soil moisture level. Moisture in the topsoil layer likely plays a role in helping I. scapularis to avoid desiccation (Lippi, Gaff, White, St. John, et al., 2021). Both soil sand and soil clay content represent soil properties that are important in water drainage through the soil where high soil sand content has better excessive water drainage ability compared to soil with high clay content, therefore another environmental variable that will keep the topmost soil layer moist such that will prevent desiccation of I. scapularis (Guerra et al., 2002). White-tailed deer (Odocoileus virginianus) are important reproductive hosts for I. scapularis; therefore, we included a raster which mapped the presence of O. virginianus in Michigan (U.S. Geological Survey (USGS) - Gap Analysis Project (GAP), 2018). 145 Multicollinearity, when the values of two or more predictors in a model matrix are highly correlated with each other (Alin, 2010), is suggested as a possible source of bias. This is because if two variables are correlated with each other, the resulting coefficients are inflated and standard errors are larger (De Marco and Nóbrega, 2018). We tested for multicollinearity among predictors using the variance inflation factor (VIF). The VIF is a numerical index that represents the degree to which variation in one predictor can be attributed to other variables due to underlying correlation structure (Alin, 2010; Vu, Muttaqi and Agalgaonkar, 2015; Cheng et al., 2022). Large VIF values (VIF > 10) may indicate significant multicollinearity between the predictor of interest and others. Therefore, any predictor with a VIF > 10 was removed from the final model matrix prior to subsequent analyses (Cheng et al., 2022). 146 Table 5.1. A description of the environmental variables used in species distribution modeling. Variable Name Final Units Source Resoluti on Bio 1 Annual Mean Temperature 1 km2 ºC WorldClim data base Bio 2 Mean Diurnal Range (Mean of monthly (max temp - min temp)) 1 km2 ºC WorldClim data base Bio 3 Isothermality (BIO2/BIO7) (×100) 1 km2 % WorldClim data base Bio 4 Temperature Seasonality (standard deviation ×100) 1 km2 % WorldClim data base Bio 5 Max Temperature of Warmest Month 1 km2 ºC WorldClim data base Bio 6 Min Temperature of Coldest Month 1 km2 ºC WorldClim data base Bio 7 Temperature Annual Range (BIO5-BIO6) 1 km2 ºC WorldClim data base Bio 8 Mean Temperature of Wettest Quarter 1 km2 ºC WorldClim data base Bio 9 Mean Temperature of Driest Quarter 1 km2 ºC WorldClim data base Bio 10 Mean Temperature of Warmest Quarter 1 km2 ºC WorldClim data base Bio 11 Mean Temperature of Coldest Quarter 1 km2 ºC WorldClim data base Bio 12 Annual Precipitation 1 km2 mm WorldClim data base Bio 13 Precipitation of Wettest Month 1 km2 mm WorldClim data base Bio 14 Precipitation of Driest Month 1 km2 mm WorldClim data base Bio 15 Precipitation Seasonality (Coefficient of Variation) 1 km2 % WorldClim data base Bio 16 Precipitation of Wettest Quarter 1 km2 WorldClim data base Bio 17 Precipitation of Driest Quarter 1 km2 mm WorldClim data base Bio 18 Precipitation of Warmest Quarter 1 km2 mm WorldClim data base Bio 19 Precipitation of Coldest Quarter 1 km2 mm WorldClim data base Snow water The average of the daily snow water equivalent (the amount of 1 km2 kg m-2 Daymet Database equivalent water contained within the snowpack) Tmax Maximum temperature 1 km2 ºC Daymet Database Tmin Minimum temperature ºC Daymet Database Elevation SRTM (Shuttle Radar Topography Mission) elevation data 1 km2 m WorldClim data base Forest groups 6 forest group layers 1 km2 NA USDA-FS 147 Table 5.1 (cont’d) Land cover Land Cover classes in 2019 put into 7 raster layers 1 km2 NA U.S. Geological classes Survey (USGS) and MRLCR Soil clay Clay content (0-2 micrometer) mass fraction in % at a depth of 1 km2 mass ISRIC World soil content 0 - 5cm. fraction in information ‰ Soil water Derived available soil water capacity (volumetric fraction) until 1 km2 volumetric ISRIC World soil retention wilting point at of soil ground observations at a dept of 0cm fraction information White-tailed Habitat distribution map for white-tailed deer (Odocoileus 1 km2 presence- USGS Gap Analysis deer presence virginianus) based on 2001 ground conditions absence 148 Species distribution models: models and evaluation Two modeling methods were used and compared to develop a species distribution model (SDM) and predict suitable habitats for I. scapularis within Michigan. The first approach used a generalized linear model (GLM) with a logit link function. The GLM method incorporates information on both the presence and absence of I. scapularis, so both were included in the data set. Presence points were defined as any sample site where I. scapularis was detected at least once during the study period, and absence points were defined as any sample site where I. scapularis was not detected over the 5-year study period. We originally used different criteria to classify presence points, ranging from the most lenient (single detection at least once during the study period) to more stringent classifications based on CDC establishment criteria (i.e., two or more ticks of any life stage or two life stages collected during a single sampling year, Dennis et al., 1998; Eisen, Eisen and Beard, 2016). We ultimately chose to use simple presence/absence criteria since more stringent criteria did not improve model performance in pilot analyses and we did not have enough data points in the Upper Peninsula (Appendix, Table 5.6). For our second approach we used a machine learning method, maximum entropy modeling (MaxEnt). Unlike the GLM, MaxEnt only requires presence points. We generated 10,000 randomly selected pseudoabsence points to extract background environmental information for MaxEnt model comparisons. The SDM package version 1.1-8 (Naimi and Araújo, 2016) in RStudio were used for GLM analyses, while the MaxEnt analysis was performed using the standalone MaxEnt application (version 3.4.4) in Java (version 8) using the provided graphical user interface (Larson et al., 2022; Phillips and Dudik Miroslav, 2008). Graphical output (e.g., response curves) for the GLM was produced on RStudio, using The SDM package version 1.1-8, while for the MaxEnt produced by the MaxEnt 149 application. The habitat suitability maps for GLM were produced by The SDM package version 1.1-8 while the habitat suitability maps for MaxEnt was produced by the MaxEnt application. Because sampling locations were unevenly distributed (Figure 3.1), sampling points were thinned to reduce the possibility of spatial pseudo replication in areas where samples were clustered (within 10 km of each other) prior to model fitting. In cases where thinning was necessary, we used the spThin package version 0.2.0 on R (Aiello-Lammens et al., 2015) to randomly select a single point within clusters that maintained >10 km distance to the next point. For the GLM, environmental data from rasters were extracted using a 10 km radial buffer around each of the sampling points using the raster package version 3.6-23 (Hijmans and van Etten, 2012) on RStudio. We then took the average of each environmental data within that 10 km radial buffer. For the categorical variables such as presence of forest classes and presence of white-tailed deer we took the average value within a 10 km radial buffer which made these variables continuous. For MaxEnt, raster values were extracted from the cell that the sampling point overlaid by the program itself. For both models, we randomly split the data into two subsets, where 70% of the data were included in the training dataset, while 30% of the data were included in the test dataset. Variable selection was conducted using two separate approaches, depending on the model. For the GLM, a bi-directional stepwise regression analysis was run independently for both Lower and Upper Peninsula analyses prior to fitting the final models for selecting variables that explained the variation seen in the presence of I. scapularis. For MaxEnt, variable selection was based on the permutation of importance to evaluate which variables were important in developing the model (Phillips and Dudik Miroslav, 2008; Larson et al., 2022). Permutation of importance can be defined as how the model performs (typically area under the curve values) in the presence of an 150 environment predictor and in the absence of that particular predictor; the higher the permutation of importance value the greater importance the variable has in explaining the variation in the data. The predictive performance of both models was evaluated using a repeated K-10-fold cross validation procedure with 10,000 iterations, where for each fold, we randomly split the data into two subsets, where 70% of the data were included in the training dataset, and 30% of the data were included in the test dataset. Model performance was evaluated for both methods based on a threshold dependent criterion which is based on the prevalence of presence points and a threshold independent criterion (Fielding and Bell, 1997b). For a threshold-independent method, the receiver operator curve (ROC) area under the curve (AUC) criteria were used. The ROC considers the true positives (sensitivity) and the false negatives. And the AUC range from 0 – 1 where a value below 0.5 indicates that the presence and absence points are distributed at random by chance and the model could not distinguish between the actual presence points and the false positives, while a value greater than 0.5 indicates the model can distinguish between a true positive and a false positive. Threshold-dependent evaluation methods included the true skill statistic (TSS) and model deviance. The TSS of a model represents the factors of the true positive rates (sensitivity) and the true negative rates (specificity) in a model (Somodi, Lepesi and Botta-Dukát, 2017). The values for TSS range from 0 – 1, where values closer to 0 represent models with low sensitivity and specificity, while a values closer to 1 represent high sensitivity and specificity. The deviance value represents the extent of the deviation of the test data from the training data. 151 RESULTS Distribution of I. scapularis in Michigan Our dataset comprised 315 sampling sites throughout Michigan (Figure 5.1) from 2017 to 2021. Sampling effort differed due to the specific surveillance objectives of each year (for a summary see Fowler et al., 2022). In 2017, sampling was focused in the southeast and the “Thumb” region of the Lower Peninsula adjacent to Lake Huron (N = 73 sites; Appendix, Figure 5.9). In 2018 (N = 153 sites), 2019 (N = 95 sites) and 2021 (N = 171 sites), sampling was carried out throughout Michigan and ranged from sites in areas with known established populations of I. scapularis to those with no previous record of occurrence (Appendix, Figure 5.9). In 2019 sites were fewer and less dense compared to that in 2018 and 2021 as sites were sampled at least three times, an objective was to obtain a better estimate of nymphal abundance as well as of prevalence of infection with the Lyme disease bacterium. Due to COVID-19 restrictions in 2020, sampling was limited to southern Michigan (N = 81 sites, Appendix, Figure 5.9). Over the 5 years, I. scapularis was detected in 175 sites (Figure 5.1), where 117 sites had established populations in at least one year according to the CDC criteria (Table 5.2). Out of the 83 counties in Michigan, I. scapularis was detected in 70 counties, where 63 counties had established populations in at least one year (Appendix, Figure 5.10). 152 Table 5.2. Sites and counties sampled in Michigan from 2017 to 2021 and the status of Ixodes scapularis presence classified as reported or established using the CDC criteria for that year (unlike the CDC, the status of Ixodes scapularis populations was determined by the data collected that year and was not contingent upon prior years’ categorizations). Year Number Number Number of sites Number of sites of sites of (Number of counties) (counties) in which I. sampled counties in which I. scapularis is scapularis is classified sampled classified as “established” as “reported” 2017 74 46 16 (15) 10 (6) 2018 180 79 37 (34) 28 (11) 2019 95 63 57 (44) 18 (11) 2020 80 27 38 (18) 23 (7) 2021 171 83 79 (53) 30 (10) 153 Figure 5.1. Sample sites in Michigan where Ixodes scapularis was detected by drag sampling in at least one year from 2017 to 2022. Environmental variables of importance Out of the 29 initial environmental predictors, 16 and 19 predictors had a VIF > 10 for the Lower Peninsula and Upper Peninsula respectively, indicating that several predictors were highly correlated. The variables that were highly correlated for both peninsulas were BIO 1, BIO 2, BIO 5, BIO 6, BIO 7, BIO 10, BIO 11, BIO 12, BIO 15, BIO 19, growing degree days, soil sand content, soil organic carbon density, open water, and forests. Furthermore, for the Lower Peninsula, soil water retention was highly correlated with the other variables mentioned above, while for the 154 Upper Peninsula, BIO 18, soil clay content, barren land and oak and hickory forest groups were highly correlated too. Thus, these variables were removed from subsequent analyses for their respective regions. Generalized linear model outputs Bidirectional stepwise regression retained the following predictors in the Lower Peninsula GLM: BIO 3, BIO 4, BIO 8, snow water equivalent, elevation, soil water retention at wilting capacity, the presence of white, red and jack pine forest group, and the presence of white-tailed deer. For the Upper Peninsula, the following predictors were retained in the GLM: BIO 3, BIO 4, BIO 13, and BIO 14, elevation, soil water retention at wilting capacity, the presence of elm, ash and cottonwood forest group, the presence of maple, beech, and birch forest group, the presence of developed land, the presence of crop land, and the presence of wetland. (Table 5.3) 155 Table 5.3. The stepwise regression coefficients for the environmental predictors for the occurrence of Ixodes scapularis that were retained in the GLM for each peninsula. See Table 1 in Methods for a list and description of all variables. Standard Region Environmental Predictors Coefficient z-value p-value Error Intercept 30.06 15.91 1.89 0.06 BIO 3 0.47 0.24 2.01 0.04 BIO 4 -0.05 0.02 -2.86 <0.01 BIO 8 0.48 0.14 3.41 <0.01 Lower Peninsula Snow water equivalent -0.09 0.05 -1.98 0.05 Elevation -0.01 0.01 -3.13 <0.01 Soil water retention at wilting capacity 0.01 0.002 3.47 <0.01 White, red and jack pine forest group presence -1.48 0.96 -1.55 0.12 White-tailed deer presence 1.98 0.69 2.87 <0.01 Intercept -81.02 34.58 -2.34 0.02 BIO 3 -1.80 0.98 -1.85 0.06 BIO 4 0.14 0.05 2.78 0.01 BIO 13 -0.26 0.14 -1.80 0.07 BIO 14 -0.28 0.18 -1.58 0.11 Elevation -0.03 0.02 -1.98 0.05 Soil water retention at wilting capacity 0.05 0.02 2.44 0.01 Elm, ash, and cottonwood forest groups presence -52.72 9140.48 -0.01 1.00 Maple, beech, and birch forest groups presence 3.05 1.82 1.68 0.09 Upper Peninsula Developed land presence -8.15 7.08 -1.15 0.25 Crop land presence -21.77 4475.37 -0.01 1.00 Wet land presence -4.43 2.85 -1.55 0.12 156 Looking at the percent contribution of each variable to the Lower Peninsula GLM, BIO 8 and snow water equivalent were two climatic predictors that contributed the most (~44%), and soil water retention at wilting capacity and elevation were the ecological predictors that contributed the most (~46%) (Table 5.4). For the Upper Peninsula GLM, BIO 3, BIO 4, and BIO 13 were the climatic predictors that contributed the most (~44%), and elevation, the presence of maple, birch and beech forest group and the presence of developed land were the ecological predictors that contributed the most (~41%) (Table 5.4) Table 5.4. Permutation of importance of environmental variables for the occurrence of Ixodes scapularis that were used in each the GLM for each peninsula. (See Table 5.1 in Methods for a list and description of all variables). Region Environmental Predictor % Contribution Soil water retention at wilting capacity 25.3 Snow water equivalent 24.7 Elevation 21.1 BIO 8 19.0 Lower Peninsula BIO 4 4.3 BIO 3 3.9 White, red and jack pine 1.6 White-tailed deer 0.1 BIO 3 18.3 BIO 14 17.2 Elevation 15.8 Maple, beech, and birch forest groups 15.3 Upper Peninsula Developed land 10.2 Wet land 9.5 BIO 4 8.6 BIO 13 3.7 Soil water retention at wilting capacity 1.4 157 In the Lower Peninsula, BIO 3 and BIO 8 have a positive association, while BIO 4 and the snow water equivalent have a negative association with the likelihood of I. scapularis occurrence Figure 5.2. The response curves for the climatic predictors of the GLM for the Lower Peninsula for the occurrence of Ixodes scapularis. The light grey shade around each curve represents the standard deviations. See Table 1 in Methods for a list and description of all variables. 158 (Figure 5.2). In the Upper Peninsula BIO 14 has a positive association; BIO 3 has stable interaction and then it decreases and increases; while BIO 4 have BIO 13 have a negative association with the likelihood of I scapularis occurrence (Figure 5.3). For the ecological predictors, in the Lower Peninsula elevation and the presence of white, red and jack pine have a negative association, while the soil water content at wilting capacity and the presence of white-tailed deer has a positive association with the likelihood of I. scapularis occurrence (Figure 5.4). In the Upper Peninsula, elevation and the presence of developed land have a negative association, while soil water content at wilting capacity, the presence of maple, birch and beech forest group and the presence of wetlands have a positive association with the likelihood of I. scapularis occurrence (Figure 5.5). 159 Figure 5.3. The response curves for the climatic predictors of the GLM for the Upper Peninsula for the occurrence of Ixodes scapularis. The light grey shade around each curve represents the +/- 1 standard deviations. See Table 1 in Methods for a list and description of all variables. 160 Figure 5.4. The response curves of the ecological predictions of the GLM for the Lower Peninsula for the occurrence of Ixodes scapularis. The gray shaded area around each curve represents the +/- 1 standard deviation. 161 Figure 5.5. The response curves of the ecological predictions of the GLM for the Upper Peninsula. The gray shaded area around each curve represents the +/- 1 standard deviation. 162 The AUC values for the Lower and Upper Peninsula GLMs are 0.78 and 0.85 respectively (Table 5.5). The TSS values for the Lower and Upper Peninsula models are 0.41 and 0.63 respectively (Table 5.5). The model deviance in the GLM was greater for the Upper Peninsula (deviance = 8.55, Table 5.5) compared to the Lower Peninsula (deviance = 1.18, Table 5.5). Table 5.5. Evaluation metrics for GLM and MaxEnt models for Ixodes scapularis presence data for the Lower and Upper Peninsulas. (AUC = Area Under the Curve; TSS = True Skill Statistic). Peninsula Model Type Evaluation criteria Value Lower AUC 0.78 GLM TSS 0.41 Model deviance 1.18 AUC 0.93 MaxEnt TSS 0.45 Model deviance 0.67 Upper AUC 0.85 GLM TSS 0.64 Model deviance 8.55 AUC 0.93 MaxEnt TSS 0.35 Model deviance 0.67 For the GLM model for the Lower Peninsula, suitable habitats for the occurrence of I. scapularis are predicted in a majority of the southwest; patchy areas in the south; and coastal areas in the southwest, southeast in the Thumb, and northeast (Figure 5.6). In contrast, the inland regions in the northwest and north central areas are estimated to be unsuitable by the GLM for the occurrence of I. scapularis (Figure 5.6). For the GLM model for the Upper Peninsula, a large portion of the southern mid region (bordering Wisconsin and Lake Michigan, including Menominee County) is estimated to be highly suitable for I. scapularis. Pockets of moderately suitable regions occur mainly in the western portion of the Lower Peninsula, while three large 163 Figure 5.6. The raster outputs of the GLM and the MaxEnt Model for Ixodes scapularis presence data for the Lower and Upper Peninsula separately. sections in the central inland and eastern regions are predicted to be least suitable for I. scapularis (Figure 5.6) MaxEnt model outputs In the MaxEnt model for the Lower Peninsula, the climatic variables BIO 9, BIO 3, BIO 8, BIO 13, and the snow water equivalent contribute the most (~48%). The ecological variables 164 elevation, presence of crop land, soil water retention at wilting capacity and the presence of white- tailed deer contributes the most (~33.2%) (Table 5.6). Table 5.6. Percent permutation of importance of environmental predictors that were retained in the MaxEnt Model for the Lower and Upper Peninsulas. See Table 1 in Methods for a list and description of all variables. % Region Environmental Predictors Contribution BIO 9 13.3 Elevation 13.9 BIO 3 10.6 BIO 8 10.3 BIO 13 8.0 Percentage of crop land 7.4 Soil water retention at wilting capacity 6.8 Snow water equivalent 5.5 Presence of white-tailed deer 5.1 BIO 4 4.7 Lower Peninsula BIO 18 4.3 BIO 14 2.5 Presence of wetland 2.2 Presence of spruce and fir forest groups 2.1 Presence of aspen and birch forest groups 1.1 Presence of grass and shrub land 0.8 Presence of developed land 0.7 Presence of white, red and jack pine forest groups 0.3 Presence of barren land 0.2 Presence of oak and hickory forest groups 0.1 Soil clay content 42.1 Elevation 10.7 Snow water equivalent 9.7 Presence of maple, birch, and beech forest groups 5.8 BIO 3 4.8 BIO 13 4.8 Upper Peninsula Presence of wetland 4.6 BIO 4 3.4 BIO 8 3.1 Soil water retention at wilting capacity 2.8 BIO 14 2.6 BIO 9 1.6 165 Table 5.1 (cont’d) Presence of white, red and jack pine forest groups 1.1 Presence of spruce and fir forest groups 1.0 Presence of aspen and birch forest groups 0.8 Presence of white-tailed deer 0.5 Presence of grass and shrub land 0.4 Presence of developed land 0.2 In the MaxEnt model for the Upper Peninsula, the climatic variables did not contribute a large amount to the model, whereas the ecological variables elevation, soil clay content, and the presence of maple, birch and beech forest group contribute the most (~59%) (Table 5.5). In the Lower Peninsula MaxEnt model, BIO 3 and BIO 8 had a positive association, while BIO 9, BIO 13 and the snow water equivalent had a negative association with the log likelihood occurrence of I. scapularis (Figure 5.7). For the ecological predictors, elevation and presence of crop land had a negative association, while the soil water content until wilting capacity and presence of white- tailed deer had a positive association with the log likelihood occurrence of I. scapularis (Figure 5.8). In the Upper Peninsula MaxEnt model, elevation and soil clay content had a negative association, while the presence of maple, birch and beech forest group had a positive association with the log likelihood occurrence of I. scapularis (Figure 5.8). Considering model evaluation, the AUC values for the Upper and Lower Peninsula both are about 0.93; the TSS values are 0.45 and 0.35, respectively (Table 5.5); and the model deviance values for the Upper and Lower Peninsula were 0.67 and 0.67, respectively. 166 Figure 5.7. The response curves for the climatic predictors of the MaxEnt model for the occurrence of Ixodes scapularis for the Lower Peninsula. The blue shaded area around each curve represents the +/- 1 standard deviations. See Table 1 in Methods for a list and description of all variables. 167 The MaxEnt model raster outputs of the suitable habitats for I. scapularis (Figure 5.6) suggest that in the Lower Peninsula, there are pockets of moderate to high levels of habitat suitability mainly in southern region, spanning from Lake Michigan in the west to Lake Huron in the east. A coastal band of moderately suitable habitat of variable width lies along Lake Michigan in the west and along Lake Huron in the east, with some small highly suitable areas dotting the coast in northwestern and northeastern regions. The mid to low suitability habitats for I. scapularis are found mainly in the northwestern and north central interior regions. In the Upper Peninsula, the areas of moderate to high suitability are in the south-central region (bordering Wisconsin and Lake Michigan, Menominee County), a few pockets of areas in the west, and a few areas along the coast of Lake Superior along the northern central region (Figure 5.6). Areas estimated to have lowest suitability lie in the far western and eastern regions, and areas with low suitability lie in the interior regions spanning the peninsula (Figure 5.6). 168 Figure 5.8. The response curves of the continuous ecological predictors of the MaxEnt models for the Lower and Upper Peninsulas. The blue shaded area around each curve represents the standard deviations. No graphs are shown for the binary predictors. 169 DISCUSSION For more than half a century I. scapularis has been expanding in the north central, northeastern, and mid-Atlantic regions in the eastern U.S. and adjacent areas in southern Canada (Dennis et al., 1998; Eisen, Eisen and Beard, 2016; Fleshman et al., 2021). Due to this range expansion of I. scapularis, the incidence of several tickborne diseases such as Lyme disease, anaplasmosis, and babesiosis have been increasing and expanding throughout the eastern U.S. as well (Bacon, Kugeler and Mead, 2008; Dahlgren et al., 2015; Kugeler et al., 2015; Fleshman et al., 2021). Michigan has been at one of the leading edges of the I. scapularis range expansion since the first established tick populations were discovered in the Upper Peninsula in the late 1980s (Strand, Walker and Merritt, 1992; EDWARD D Walker et al., 1994) and in the Lower Peninsula (Erik Scott Foster, 2004) in early 2000s. Following the discovery of these populations, I. scapularis has expanded in the central and western regions of the Upper Peninsula as well as in the southern and coastal regions in the Lower Peninsula (Hamer et al., 2014; Lantos et al., 2017; Fleshman et al., 2021). Given I. scapularis is still emerging in Michigan, we wanted to identify areas in Michigan where I. scapularis may have a high likelihood of invading and becoming established. To do so, we used two methods of species distribution modeling, one was a regression-based method (a generalized linear model), and the other was a machine learning method (MaxEnt) to identify the environmental predictors and the suitable habitats in Michigan for I. scapularis occurrence. Why model the two peninsulas separately? The 11th largest state in the U.S., Michigan is comprised of a southern and northern peninsula that differ greatly in climatic, habitat and soil conditions. Given differences in geography (Omernik and Griffith, 2014) as well as invasion history, we modeled the two peninsulas 170 separately. Furthermore, we had many more sampling sites (and more positive sites) in the Lower Peninsula compared to the Upper Peninsula (Figure 5.1), and the model would be weighted towards the conditions in the Lower Peninsula. Altogether from 2017 to 2021, we had 222 and 60 (21.3%) sites in the Lower and Upper Peninsulas respectively, from which I. scapularis was detected in 124 and 27 (17.9%) sites respectively. We did in fact explore how different model outputs would be when both peninsulas were modelled together, using both GLM and MaxEnt modeling approaches (Appendix, Table 5.6). Differences in model outputs (between when modeling peninsulas together versus separately) were more pronounced for the Upper Peninsula (Appendix, Figure 5.11). When modeled together, only one small portion of the Upper Peninsula seemed to have suitable habitats for I. scapularis (Appendix, Figure 5.11), whereas when modeled separately, more suitable habitats were defined, including several hotspots in the western portion of the state where established populations of I. scapularis had been detected previously. We did not observe a drastic difference in the distribution of suitable habitats for I. scapularis when the Lower Peninsula was modeled together or separately, perhaps owing to the larger number of sites at which I. scapularis had been detected throughout much of the peninsula. The GLM and MaxEnt model results were generally in agreement with each other for both the Lower Peninsula and the Upper Peninsula for estimating both highly suitable regions and unsuitable habitats. The Lower Peninsula The first established population in the Lower Peninsula was discovered in the southwest corner of the state in the early 2000s (Foster 2004), potentially invading from neighboring states of Indiana and Illinois where I. scapularis was established previously and found questing and on host mammals (Pinger, Timmons and Karris, 1996; Jones and Kitron, 2000). Ixodes scapularis continued to spread, initially faster northwards along Lake Michigan than inland and eastwards 171 (Hamer et al., 2014), but has since become detected and/or established 52 out of 83 counties (Appendix, Figure 5.10) (Dennis et al., 1998; Eisen, Eisen and Beard, 2016; Lantos et al., 2017; Fleshman et al., 2021). Despite becoming more widespread in the Lower Peninsula, there are still many areas and counties where I. scapularis has not yet invaded. The climatic predictors isothermality (BIO 3), mean temperature of the wettest quarter (BIO 8) and snow water equivalent were the three climatic predictors found in common to both the GLM and the MaxEnt in contributing substantially to explaining the variation seen in I. scapularis presence sites. Isothermality is a measure of how differences in day to night temperatures vary in relation to differences in summer to winter temperatures (Kessler, Ganser and Glass, 2019). If the isothermality is high, it relates to temperature uniformity while low isothermality relates to greater temperature variation. In both models there was a positive association with isothermality and I. scapularis occurrence. In the Lower Peninsula, isothermality is generally high in most of the southern regions while in the Upper Peninsula there is lower isothermality on the eastern region compared to the west. Especially given on average higher temperatures in southern Michigan, having more uniform temperatures is important for successful development of I. scapularis where it ensures that there are enough degree days for I. scapularis to develop to its next life stage. A similar relationship is seen in Dermacentor variabilis, the American dog tick, where it is hypothesized that higher isothermality may aid in developmental rates (James et al., 2015; Kessler, Ganser and Glass, 2019). The temperature of the wettest quarter has a positive relationship for likelihood of I. scapularis occurrence. For Michigan the wettest quarter of the year is April – June which overlaps with the adult and nymphal host-seeking activity periods of I. scapularis. An increase in temperature during this quarter (when relative humidity should also be conducive to host-seeking 172 and survivorship), may lead to increased questing activity of I. scapularis, which theoretically should allow ticks greater chances for finding hosts (Vail and Smith, 1998; Valsson and Bharat, 2011; Berger et al., 2014; Elias et al., 2021). The snow water equivalent relates to the snowpack density. When the snowpack density is higher, it provides an insulation from the lower air temperatures during the harsh winter months and increase I. scapularis overwintering success (Volk et al., 2022). Interestingly our model showed a negative association between snow water equivalent and the likelihood of I. scapularis occurrence (Figure 5.2). One possibility in the Lower Peninsula is the snow water equivalent shows to be the greatest in the northern region where we did not detect any I. scapularis, there by our model predicts this negative association. For the GLM, the temperature seasonality (BIO 4) had a negative association with I. scapularis occurrence. When the temperature variation becomes greater, how different winter and summer temperatures are may impact the developmental patterns and the activity patterns of I. scapularis resulting in lower occurrence in regions where there is a greater temperature variation (Vail and Smith, 1998; Schulze, Jordan and Hung, 2001; Burtis et al., 2016). In the Lower Peninsula there is a higher temperature variation inland compared to coastal regions, and within the Upper Peninsula, there is a higher temperature variation in west compared to the east. Additionally, for the MaxEnt model average temperature of driest quarter (BIO 9) and the amount of precipitation of in the wettest month (BIO 13) were important climatic predictors. The driest quarter of the year in Michigan falls in January – March. In general, there is a positive association with BIO 9 and the likelihood of I. scapularis occurrence which corresponds to milder winter temperatures, which may result in higher adult I. scapularis host-seeking activity (Valsson 173 and Bharat, 2011; Burtis et al., 2016; Ogden and Lindsay, 2016; Bouchard et al., 2019; Wallace et al., 2019). The other climatic variable is the precipitation of wettest month (BIO13), which in Michigan, falls around April. There was a negative association with the amount of precipitation of the wettest month and log likelihood of I. scapularis occurrence. High levels of precipitation may create either flooding or occurrence of pathogens like fungal infections on I. scapularis which would be unsuitable conditions (Berger et al., 2014; Ogden and Lindsay, 2016; Bouchard et al., 2019). Considering the ecological predictors, elevation was an important predictor in both the GLM and MaxEnt models, where it was negatively associated with the likelihood of I. scapularis occurrence. This finding is consistent with other studies, where it is hypothesized that conditions (e.g., too dry, too cold, too low host abundance) decrease I. scapularis survivorship (Diuk-Wasser et al., 2010; Hahn et al., 2016). In these studies, however, elevation differences were referring to elevations above 500 – 800 m, such as the Appalachians. The elevational gain in Michigan, however, is substantially less, especially in the Lower Peninsula, and the change in the elevation may be a proxy for moving from the coast to inland regions. There may be climatic and ecological reasons underlying this association, but it may also be an artifact of the invasion of I. scapularis which has spread initially from coastal, i.e., low land areas, within the state. One way to test this hypothesis is modeling the coastal regions and testing it on the inlands regions to see if elevation or the proximity to the coast influences the likelihood of occurrence of I. scapularis. Soil water content at wilting capacity was another variable that was important in both the GLM and the MaxEnt models for the Lower Peninsula. Soil water retention ability correlates to the moisture content in the soil which will keep the soil layers humid such that it will be helpful 174 for the survival of I. scapularis (Schulze, Jordan and Hung, 2001; Rodgers, Zolnik and Mather, 2007; Hayes, Scott and Stafford, 2015; Burtis and Pflueger, 2017; Ripoche et al., 2018; Larson et al., 2022). In the GLM, the presence of white-tailed deer was positively associated with the likelihood of I. scapularis. White-tailed deer are important reproductive hosts for the adult stage of I. scapularis. Thus, the presence of deer is important for the survival of I. scapularis, and other models have shown that the presence of white-tailed deer is important to the distribution of I. scapularis (Elias et al., 2021; Kopsco et al., 2023). Conversely, in the GLM, there was a negative association with the presence of white, red and jack pine forests groups and I. scapularis occurrence. Pine forests tend to have a thinner layer of leaf litter; be drier; and have more acidic clay soil (Guerra et al., 2002; Lubelczyk et al., 2004), which may be unfavorable for the survival of I. scapularis. Similar to elevation, however, it also could be that areas where pines predominate (e.g., central northern Michigan) are those where I. scapularis has not yet reached. Alternatively, this finding may also reflect a bias in our sampling design. In our surveillance, we mainly focused on the sites that had deciduous woods or mixed deciduous and coniferous woods, because those are the habitats with sufficient leaf litter that should facilitate survival of I. scapularis (Ginsberg, Rulison, Miller, Pang, Arsnoe, Hickling, Ogden, LeBrun, et al., 2020). Because the GLM and MaxEnt models use different sets of assumptions and different modeling methods, the AUC, TSS and model deviance values are not completely comparable. Although overfitting was not observed in either model, MaxEnt uses machine learning techniques which will maximize the pattern matching of the presence points, which will result in greater AUC values compared to the GLM. In our models the TSS ranged between 0.4 – 0.6 indicating relatively our test data set was able to predict the model based on the training data set well. 175 The Upper Peninsula The first established populations of I. scapularis discovered in Michigan were discovered in Menominee County, which is in the southern central region of the Upper Peninsula bordering Wisconsin and Lake Michigan (Strand, Walker and Merritt, 1992; Walker et al., 1998). In the final GLM model for the Upper Peninsula, there was one unique climatic predictor that was not present for the Lower Peninsula model- precipitation of the driest month (BIO14) has a positive association with I. scapularis occurrence. The driest month in Michigan typically is February, during which time I. scapularis generally is expected to be inactive. During winter, larvae, and nymphs most likely are in diapause (Ogden et al., 2018), and only adult ticks can quest, but only when temperatures are above freezing. Thus, at this time of year in northern Michigan, increased precipitation likely would-be increased snow, which might help tick overwintering survivorship. High levels of precipitation keep the topmost soil layer and bottom of the leaf litter layer warm and humid which would be conducive for the survival of I. scapularis. In the Upper Peninsula, the amount of temperature variation over a given period (BIO 4) has a positive association while in the Lower Peninsula it has a negative association. It is possible that great variation indicates that the summers are warm enough, and potentially long enough, to allow for more successful host-seeking success and developmental success from one life stage to another. Colder temperatures are not as much of a concern for I. scapularis if ticks are protected under leaflitter and snow, as written previously. Interestingly none of the bioclimmatic predictors for the Upper Peninsula of the MaxEnt model were important in explaining the variation. This may be due to the variation in bioclimmatic predictors not being drastically different among landscape for the MaxEnt model to pick up a drastic difference. 176 Considering the ecological predictors unique in the GLM, presence of maple beech birch forest group and presence of wetlands have positive associations. The positive association with deciduous forests is not surprising given what is known about I. scapularis ecology (Randolph, 2004; Pfäffle et al., 2013), what others have previously found (Brownstein, Holford and Fish, 2003; Diuk-Wasser, Gatewood, et al., 2006; Diuk-Wasser et al., 2010) and may reflect our sample site selection bias. What is surprising, however, is the positive association between the presence of wetlands and I. scapularis presence. In general wetlands can be too wet or moist for I. scapularis (Larson et al., 2022). There may be regions in the Upper Peninsula, however, which are not covered in water/ice the whole year especially wooded wetlands which might be good habitats for I. scapularis. Soil clay content is negatively associated with I. scapularis occurrence. Soil with larger clay content is poorly drained, and as mentioned above, too high moisture and water content in the soil is unfavorable for I. scapularis. Evaluating the model performance for the Upper Peninsula again AUC values were > 0.5 for both model approaches, indicating the model can categorize the presence sites and absence sites accurately than by random. It is important to note that the model deviance for the Upper Peninsula in the GLM was high (8.6). Deviance values range from 0 – infinity and smaller values indicate a better fit of the model to the data. We had a relatively smaller sample size of sites and of positive sites for the Upper Peninsula, and when 30% of the data was held back from training in order to be used to test the model, we had so few sites, which may have caused the deviation of model to be greater. To test this hypothesis, future research could add more sites and sample more frequently targeting the Upper Peninsula. One can also artificially randomly reduce the dataset from the Lower Peninsula, the models and see how deviance changes with sample size. It is interesting to note that the deviance values for the MaxEnt for both the two Peninsula were similar, 177 and this could be partly due to an artifact of the modeling method where the MaxEnt being a machine learning method would maximize pattern recognition among the presence points and decrease the model deviance. Species distribution models: comparisons with published models Using two approaches, we modeled potential suitable habitats for I. scapularis in Michigan based on the surveillance data collected 2017 to 2021. Because I. scapularis is continuing to spread in Michigan, it is unclear how well the maps produced based on a five-year snapshot of data will resemble the future distribution of ticks, such as in another two decades. The GLM and MaxEnt modeling processes are pattern matching approaches – one using regression and one using machine learning – that then assign probabilities of I. scapularis occurrence based on relationships identified between the data (i.e., tick presence/absence at various sites) and parameters (i.e., environmental variables) provided. There are many reasons to “believe” in the reliability of these models, such as examining the evaluating criteria mentioned previously like AUC, TSS and model deviance and looking at how well the models predict the known presence sites. There are also reasons to believe that the modeled habitat suitability maps can be improved, such as adding in more presence absence points, using one year’s presence-absence points, and then projecting onto the next years presence-absence points to see how well they overlap and projecting a state or an endemic region onto a tick expansion region. To be specific, however, we have modeled the habitat suitability of dynamic I. scapularis populations as they invade across Michigan landscapes, and not necessarily the habitat suitability of more equilibria I. scapularis populations that have reached an endemic state. The latter would obviously produce a more accurate model, but from the standpoint of making predictions in the nearer term that can still be helpful for public health and our understanding of the invasion process, our models still can be useful. 178 With time as I. scapularis populations expand, potentially most of the state- especially in the Lower Peninsula- might be deemed as having suitable habitats in the future. In distribution models presented by Burtis et al., 2022, they used county level data on I. scapularis presence/absence that are available for the eastern U.S. and then modelled the potential suitable habitats at a county scale. According to their models all Michigan counties are predicted to have suitable habitats. The wide range of environmental conditions present in other I. scapularis- endemic areas generally must have encompassed that found in Michigan and resulted in all counties estimated to have suitable habitats. In another model developed in Hahn et al., 2016, which used I. scapularis distribution data from different literature sources (e.g., Eisen et al., 2016) indicated high suitability counties along the western coast and within the northern central region of the Lower Peninsula as well as almost all regions of the Upper Peninsula, while the Thumb region and the southeastern regions were less suitable. This is slightly different from what was reported in our model as well as the model by Burtis et al., 2022. This may be due to when the models were developed, data used in these models were current up through 2015 (i.e., including areas where I. scapularis is established). Data from Michigan that would have been included in analyses would have shown I. scapularis populations mainly in the southwestern and western Lower Peninsula (but not in the Thumb nor southeastern Michigan), and limited areas in western Upper Peninsula (Lantos et al. 2017). The Burtis et al. 2022, Hahn et al., 2016 models report suitability at the county level, but a similar approach, whereby one uses data from other areas where I. scapularis is already endemic, could be conducted to predict habitat suitability at a finer scale. For example, given Michigan shares much of its major environmental conditions with Wisconsin, where I. scapularis has been endemic for many years, one could develop a habitat suitability model based on Wisconsin data 179 and extrapolate onto Michigan. This is exactly what Foster 2004 did; Foster and colleagues discovered the first populations of I. scapularis in southwestern Michigan by projecting a habitat suitability model created by Guerra et al, 2002 based on data from Wisconsin onto Michigan. Those data were collected from 1996 – 1998, when similar to the current situation in Michigan, I. scapularis was still in the process of spreading in Wisconsin. The model in Guerrra et al. 2002 shows slightly different regions of suitable habitats compared to our models, which is surprising given the different training and environmental data on which the models are built. The most interesting difference is that Guerra et al., 2002 predicted the large area in the interior north central Lower Peninsula to be highly suitable, whereas in our models currently estimate them to be not suitable for I. scapularis. Indeed, our model may not be correct because I. scapularis has not yet invaded areas with similar abiotic and ecological features to that found in north central Michigan. Thus, MaxEnt may not identify those habitats as suitable, and our absence datapoints for sites in that region would strengthen any negative association in the GLM. Having said that, much of this region is characterized by moderate to low precipitation and the shortest growing seasons in Michigan (Dickmann and Leefers, 2016), and both of which reduce tick survivorship and population establishment of I. scapularis (Lindsay et al. 1995). Another model developed using I. scapularis nymphal densities is presented in Diuk-Wasser et al., 2006, 2010. These models were based on nymphal densities and during that early time because I. scapularis was only detected in the southwestern region of the Lower Peninsula, most of the southern regions were deemed to have low suitability. Limitations and future research As mentioned, previously, because Michigan is still undergoing an invasion process, one major limitation to our ability to build an accurate model for habitat suitability of I. scapularis in 180 Michigan is that the presence/absence data may continue to change over time as I. scapularis spreads to new areas. We believe our presence data is generally reliable, especially in areas that were sampled in multiple years, and given our experience and the literature that, once I. scapularis invades an area, it likely will become established (e.g., Hamer et al. 2010). Regarding our absence data, however, it is unclear whether they will remain negative in the future, again, based on our experience and trends in the literature. Even though some sites from our model may be modeled as unsuitable now, they may become suitable in the future, as predicted by other models using data from other areas in the northern eastern US. One of the reasons we did not use the same approach as others – to extrapolate models based on data from other states to Michigan - is because of Michigan’s unique geography – comprising of two relatively large landmasses in the middle of the continent but being largely surrounded by large bodies of water. Having said that, given the challenge of the on-going invasion, using models based on data from other areas makes sense. Thus, future research should include updating the Guerra et al. 2002 model based on current data from Wisconsin (or more broadly in the Upper Midwest) and projecting it onto Michigan as per Foster, 2004. Wisconsin is the closest and most environmentally similar area with Michigan, even if there are environmental differences. The model from Guerra et al., 2002 was based on presence data distributed throughout western Wisconsin with a few exceptions, including sites from northeastern WI and Menominee County (Strand, Walker and Merritt, 1992; Walker et al., 1998). In subsequent years, I. scapularis spread throughout the state (as well as into other states) and has had time to become established across the landscape. Thus, a new habitat suitability model may be able to provide an even more reliable and finer scale model. It would be interesting to test and compare predictions made by Guerra et al., 2002 on current Wisconsin data as well as projecting it onto Michigan to forecast 181 suitable habitats. Conducting a similar exercise using tick distribution data from Ontario, Canada may also be useful. Another exercise that could shed light on how models are affected by the input data, given the invasion dynamics, is to re-run our analyses (or, better, an updated Wisconsin habitat suitability model run on Wisconsin data) and intentionally train the models on non-random subsets of data to see how similar the resulting models would be. For our model evaluation, models are trained multiple times with random subsets of data to obtain outputs with measures of uncertainty. But our complete data set is influenced heavily by invasion from the southwest area of Michigan in the Lower Peninsula and south-central Upper Peninsula. How different, for example, would the model for the Lower Peninsula currently look if the initial invasion had occurred from southeastern Michigan? Or how would the habitat suitability model for Wisconsin look if presence data originated from the eastern half of the state near Lake Michigan, rather than from the western half and near the Driftless Zone? This is of interest because as I. scapularis and other species invade an area, it is important to understand for building species distribution models that current associations with habitats are affected by historical contingencies (e.g., what was the point of introduction?), the ecological context – including what habitats are available and how representative that region of invasion is of the rest of the geographic extent of interest. Future research to build potentially more precise maps in Michigan include continuing to conduct surveillance, especially in areas where I. scapularis currently is absent, as well as to model habitat suitability using abundance data (e.g., with negative binomial regression models or zero- inflated negative binomial models) and not just presence/absence data. Currently the number of sites with abundance data is too low to be able to build a reliable model, but as established 182 populations continue to increase in size and as the invasion continues, more data should become available. To test our current habitat suitability model directly and not just rely on using updated models developed in Wisconsin, or to just wait for invasion to occur in Michigan, there are two things we can do. One is to plan active surveillance targeting those potentially suitable and unsuitable regions where we have not detected the presence/establishment of I. scapularis. The second is to conduct tick survivorship and development studies of each life stage in these regions by putting them out in semi-natural arenas to measure the effects of local abiotic factors (e.g., Lindsay et al., 1998; Ginsberg et al., 2014; Ogden et al., 2018; Volk et al., 2022). Future studies can also conduct species distribution modeling on not just I. scapularis in general but on ticks infected with various disease agents. Understanding where subsets of I. scapularis ticks are infected with the agents of Lyme disease, human granulocytic anaplasmosis, babesiosis, Powassan virus encephalitis, and others may have benefits for directing public health messages regrading prevention, diagnosis, and treatment, but also better understand the variation in ecology of the tick as the pathogens have similar but different ecologies. With the increase in I. scapularis borne diseases it is important to anticipate the patterns of the spread of I. scapularis. Currently there are no accepted area-wide tick control methods, unfortunately, so prevention of tick-borne diseases relies mainly on personal protection, and therefore education of the public, public health workers, and medical workers. Species distribution models can serve as a guide for surveillance and a guide to design any control and intervention methods when they become available. 183 BIBLIOGRAPHY Adams, D.A. et al. (2015) Summary of Notifiable Infectious Diseases and Conditions — United States, 2015. Available at: https://www.cdc.gov/MMWR/. Adler, G.H., et al. (1992) ‘Vegetation structure influences the burden of immature Ixodes dammini on its main host, Peromyscus leucopus’, Parasitology, 105(1), pp. 105–110. Available at: https://doi.org/10.1017/S0031182000073741. Aiello-Lammens, M.E. et al. (2015) ‘spThin: An R package for spatial thinning of species occurrence records for use in ecological niche models’, Ecography, 38(5), pp. 541–545. Available at: https://doi.org/10.1111/ecog.01132. Alin, A. (2010) ‘Multicollinearity’, Wiley Interdisciplinary Reviews: Computational Statistics, 2(3), pp. 370–374. Available at: https://doi.org/10.1002/wics.84. Bacon, E.A. et al. (2022) ‘Effects of Climate on the Variation in Abundance of Three Tick Species in Illinois’, Journal of Medical Entomology, 59(2), pp. 700–709. Available at: https://doi.org/10.1093/jme/tjab189. Bacon, R.M., Kugeler, K.J. and Mead, P.S. (2008) ‘Surveillance for Lyme Disease-United States, 1992-2006.’, MMWR. CDC Surveillance summaries: Morbidity and mortality weekly report, pp. 1–9. Barbour, A.G. and Fish, D. (1993) ‘The Biological and Social Phenomenon of Lyme Disease’, Science, 260(5114), pp. 1610–1616. Available at: https://doi.org/10.1126/science.8503006. Berger, K.A. et al. (2014) ‘Relative humidity and activity patterns of Ixodes scapularis (Acari: Ixodidae)’, Populations and Community Ecology, 51(4), pp. 769–776. Available at: https://doi.org/https://doi.org/10.1603/ME13186. Bertrand, Matthew R and Wilson, M.L. (1996) ‘Microclimate-Dependent Survival of Unfed Adult Ixodes scapularis (Acari: Ixodidae) in Nature: Life Cycle and Study Design Implications’, J. Med. Entomol, 33(4). Available at: https://doi.org/https://doi.org/10.1093/jmedent/34.2.167. Bertrand, Matthew R. and Wilson, M.L. (1996) ‘Microclimate-Dependent Survival of Unfed Adult Ixodes scapularis (Acari: Ixodidae) in Nature: Life Cycle and Study Design Implications’, Journal of Medical Entomology, 33(4), pp. 619–627. Available at: https://doi.org/10.1093/jmedent/33.4.619. Bouchard, C. et al. (2019) ‘Increased risk of tick-borne diseases with climate and environmental changes’, Canada Communicable Disease Report, 45(4), pp. 83–89. Available at: https://doi.org/10.14745/ccdr.v45i04a02. 184 Bouseman, J.K. et al. (1990) ‘Status of Ixodes dammini (Acari: Ixodidae) in Illinois’, Journal of Medical Entomology, 27(4), pp. 556–560. Available at: https://doi.org/10.1093/jmedent/27.4.556. Brownstein, J.S., Holford, T.R. and Fish, D. (2003) ‘A climate-based model predicts the spatial distribution of the Lyme disease vector Ixodes scapularis in the United States’, Environmental Health Perspectives, 111(9), pp. 1152–1157. Available at: https://doi.org/10.1289/ehp.6052. Brunner, J.L. et al. (2023) ‘Off-host survival of blacklegged ticks in eastern North America: A multistage, multiyear, multisite study’, Ecological Monographs, e1572. Available at: https://doi.org/10.1002/ecm.1572. Bunnell, J.E. et al. (2003) ‘Geographic Information Systems and Spatial Analysis of Adult Ixodes scapularis (Acari: Ixodidae) in the Middle Atlantic Region of the U.S.A.’, Journal of Medical Entomology, 40(4), pp. 570–576. Available at: https://doi.org/10.1603/0022-2585- 40.4.570. Burtis, J.C. et al. (2016) ‘The impact of temperature and precipitation on blacklegged tick activity and Lyme disease incidence in endemic and emerging regions’, Parasites and Vectors, 9(1), pp. 1–10. Available at: https://doi.org/10.1186/s13071-016-1894-6. Burtis, J.C. and Pflueger, C. (2017) ‘Interactions between soil-dwelling arthropod predators and Ixodes scapularis under laboratory and field conditions’, Ecosphere, 8(8). Available at: https://doi.org/10.1002/ecs2.1914. Burtis, J.C. and Pflueger, C. (2017) ‘Interactions between soil-dwelling arthropod predators and Ixodes scapularis under laboratory and field conditions’, Ecosphere, 8(8). Available at: https://doi.org/10.1002/ecs2.1914. Duffy, D.C. and Campbell, S.R. (1994) ‘Ambient Air Temperature as a Predictor of Activity of Adult Ixodes scapularis (Acari: Ixodidae)’, Journal of Medical Entomology, 31(1), pp. 178–180. Available at: https://doi.org/10.1093/jmedent/31.1.178. Centers for Disease Control and Prevention (2019) ‘Surveillance for Ixodes scapularis and pathogens found in this tick species in the United States’. Centers for Disease Control and prevention. Cheng, J. et al. (2022) ‘A variable selection method based on mutual information and variance inflation factor’, Spectrochimica Acta - Part A: Molecular and Biomolecular Spectroscopy, 268. Available at: https://doi.org/10.1016/j.saa.2021.120652. Clayton, L., Attig, J.W. and Mickelson, D.M. (2001) ‘Effects of late Pleistocene permafrost on the landscape of Wisconsin, USA’, Boreas, 30(3), pp. 173–188. Available at: https://doi.org/10.1111/j.1502-3885.2001.tb01221.x. 185 Clifford, C.M., Anastos, G. and Elbl, A. (1961) ‘The larval Ixodid ticks of the Eastern United States (Acarina - Ixodidae)’, Misc. Publ. of the Entomological Society of America, 2(5), pp. 213– 237. Couper, L.I. et al. (2020) ‘Comparative vector competence of North American Lyme disease vectors’, Parasites and Vectors, 13(1). Available at: https://doi.org/10.1186/s13071-020-3893-x. Coyle, D.R. et al. (2013) ‘Belowground herbivory in red pine stands initiates a cascade that increases abundance of Lyme disease vectors’, Forest Ecology and Management, 302, pp. 354– 362. Available at: https://doi.org/10.1016/j.foreco.2013.03.017. Dahlgren, F.S. et al. (2015) ‘Human granulocytic anaplasmosis in the United States from 2008 to 2012: a summary of national surveillance data.’, The American journal of tropical medicine and hygiene, 93(1), pp. 66–72. Available at: https://doi.org/10.4269/ajtmh.15-0122. Decker, K.L.M. et al. (2003) ‘Snow Removal and Ambient Air Temperature Effects on Forest Soil Temperatures in Northern Vermont’, Soil Science Society of America Journal, 67(4), pp. 1234–1242. Available at: https://doi.org/10.2136/sssaj2003.1234. Dennis, D.T. et al. (1998) ‘Reported Distribution of Ixodes scapularis and Ixodes pacificus (Acari: Ixodidae) in the United States’, Journal of Medical Entomology, 35(5), pp. 629–638. Available at: https://doi.org/10.1093/jmedent/35.5.629. Dewitz, J. and U.S. Geological Survey (2021) ‘National Land Cover Database (NLCD) 2019 Products (ver. 2.0, June 2021)’, U.S. Geological Survey data release [Preprint]. Available at: https://doi.org/doi.org/10.5066/P9KZCM54. Dickmann, D.I. and Leefers, L.A. (2016) ‘The forests of Michigan today’, University of Michigan Press. Diuk-Wasser, M.A. et al. (2006) ‘Spatiotemporal patterns of host-seeking Ixodes scapularis nymphs (Acari: Ixodidae) in the United States’, Journal of Medical Entomology, 43(2), pp. 166– 176. Available at: https://doi.org/10.1603/0022-2585(2006)043. Diuk-Wasser, M.A. et al. (2010) ‘Field and climate-based model for predicting the density of host-seeking nymphal Ixodes scapularis, an important vector of tick-borne disease agents in the eastern United States’, Global Ecology and Biogeography, 19(4), pp. 504–514. Available at: https://doi.org/10.1111/j.1466-8238.2010.00526.x. Diuk-Wasser, M.A., Vanacker, M.C. and Fernandez, M.P. (2021) ‘Impact of Land Use Changes and Habitat Fragmentation on the Eco-epidemiology of Tick-Borne Diseases’, Journal of Medical Entomology, 58(4), pp. 1546–1564. Available at: https://doi.org/10.1093/jme/tjaa209. Durden, L.A. and Keirans, J.E. (1997) ‘Nymphs of the Genus Ixodes (Acari: Ixodidae) of the United States: Taxonomy, Identification Key, Distribution, Hosts, and Medical/Veterinary 186 Importance’, Thomas Say Publications in Entomology: Monographs. Entomological Society of America. Available at: https://doi.org/10.2307/3495570. Eisen, R.J. et al. (2016) ‘Linkages of Weather and Climate with Ixodes scapularis and Ixodes pacificus (Acari: Ixodidae), Enzootic Transmission of Borrelia burgdorferi, and Lyme Disease in North America’, Journal of Medical Entomology, 53(2), pp. 250–261. Available at: https://doi.org/10.1093/jme/tjv199. Eisen, R.J. and Eisen, L. (2018) ‘The Blacklegged Tick, Ixodes scapularis: An Increasing Public Health Concern’, Trends in Parasitology. Elsevier Ltd, pp. 295–309. Available at: https://doi.org/10.1016/j.pt.2017.12.006. Eisen, R.J., Eisen, L. and Beard, C.B. (2016) ‘County-scale distribution of Ixodes scapularis and Ixodes pacificus (Acari: Ixodidae) in the continental United States’, Journal of Medical Entomology, 53(2), pp. 349–386. Available at: https://doi.org/10.1093/jme/tjv237. Elias, S.P. et al. (2006) ‘Deer Browse Resistant Exotic-Invasive Understory: An Indicator of Elevated Human Risk of Exposure to Ixodes scapularis (Acari: Ixodidae) in Southern Coastal Maine Woodlands’, Journal of Medical Entomology, 43(6), pp. 1142–1152. Available at: https://doi.org/10.1093/jmedent/43.6.1142. Elias, S.P. et al. (2021) ‘A generalized additive model correlating blacklegged ticks with white- tailed deer density, temperature, and humidity in Maine, USA, 1990-2013’, Journal of Medical Entomology, 58(1), pp. 125–138. Available at: https://doi.org/10.1093/jme/tjaa180. Elias, S.P. et al. (2022) ‘Emergence of Ixodes scapularis (Acari: Ixodidae) in a Small Mammal Population in a Coastal Oak-Pine Forest, Maine, USA’, Journal of Medical Entomology, 59(2), pp. 725–740. Available at: https://doi.org/10.1093/jme/tjab209. Estrada-Pea, A. (2001) ‘Forecasting habitat suitability for ticks and prevention of tick-borne diseases’, Veterinary Parasitology, 98(1–3), pp. 111–132. Available at: https://doi.org/10.1016/S0304-4017(01)00426-5. Estrada-Peña, A. (2002) ‘Increasing habitat suitability in the United States for the tick that transmits Lyme disease: A remote sensing approach’, Environmental Health Perspectives, 110(7), pp. 635–640. Available at: https://doi.org/10.1289/ehp.02110635. Feria-Arroyo, T.P. et al. (2014) ‘Implications of climate change on the distribution of the tick vector Ixodes scapularis and risk for Lyme disease in the Texas-Mexico transboundary region’, Parasites and Vectors, 7(1), pp. 1–16. Available at: https://doi.org/10.1186/1756-3305-7-199. Fick, S.E. and Hijmans, R.J. (2017) ‘WorldClim 2: new 1-km spatial resolution climate surfaces for global land areas’, International Journal of Climatology, 37(12), pp. 4302–4315. Available at: https://doi.org/10.1002/joc.5086. 187 Fielding, A.H. and Bell, J.F. (1997) ‘A review of methods for the assessment of prediction errors in conservation presence/absence models’, Environmental Conservation, 24(1), pp. 38–49. Available at: https://doi.org/10.1017/S0376892997000088. Fish, D. (2021) ‘Range expansion of Ixodes scapularis in the USA’, in P. Nuttall (ed.) Climate, ticks and disease. Wallingford: CABI, pp. 176–183. Available at: https://doi.org/10.1079/9781789249637.0000. Fisher, T.G., Jol, H.M. and Boudreau, A.M. (2005) ‘Saginaw Lobe tunnel channels (Laurentide Ice Sheet) and their significance in south-central Michigan, USA’, in Quaternary Science Reviews, pp. 2375–2391. Available at: https://doi.org/10.1016/j.quascirev.2004.11.019. Fleshman, A.C. et al. (2021) ‘Reported County-Level Distribution of Lyme Disease Spirochetes, Borrelia burgdorferi sensu stricto and Borrelia mayonii (Spirochaetales: Spirochaetaceae), in Host-Seeking Ixodes scapularis and Ixodes pacificus Ticks (Acari: Ixodidae) in the Contiguous United States’, Journal of Medical Entomology. Edited by M. Diuk-Wasser, 58(3), pp. 1219– 1233. Available at: https://doi.org/10.1093/jme/tjaa283. Foster, E.S. (2004) ‘Ixodes scapularis (Acari:Ixodidae) and Borrelia burgdorferi in Southwest Michigan: Population ecology and verification of a geographic risk model.’ Michigan State University. Fowler, P.D. et al. (2022) ‘Northward Expansion of Amblyomma americanum (Acari: Ixodidae) into Southwestern Michigan’, Journal of Medical Entomology, 59(5), pp. 1646–1659. Available at: https://doi.org/10.1093/jme/tjac082. Gabriele-Rivet, V. et al. (2015) ‘Different ecological niches for ticks of public health significance in Canada’, PLoS ONE, 10(7), pp. 1–19. Available at: https://doi.org/10.1371/journal.pone.0131282. Gardner, A.M. et al. (2020) ‘Landscape features predict the current and forecast the future geographic spread of Lyme disease: Landscape predicts Lyme disease spread’, Proceedings of the Royal Society B: Biological Sciences, 287(1941). Available at: https://doi.org/10.1098/rspb.2020.2278rspb20202278. Ginsberg, H.S. et al. (2004) ‘Woodland Type and Spatial Distribution of Nymphal Ixodes scapularis (Acari: Ixodidae)’, Environmental Entomology, 33(5), pp. 1266–1273. Available at: https://doi.org/10.1603/0046-225X-33.5.1266. Ginsberg, H.S. et al. (2014) ‘Comparison of survival patterns of northern and southern genotypes of the North American tick Ixodes scapularis (Acari: Ixodidae) under northern and southern conditions’, Parasites and Vectors, 7(1), pp. 1–10. Available at: https://doi.org/10.1186/1756- 3305-7-394. Ginsberg, H.S. et al. (2017) ‘Environmental factors affecting survival of immature Ixodes scapularis and implications for geographical distribution of Lyme Disease: The climate/behavior 188 hypothesis’, PLoS ONE, 12(1), pp. 1–17. Available at: https://doi.org/10.1371/journal.pone.0168723. Ginsberg, H.S. et al. (2020) ‘Local abundance of Ixodes scapularis in forests: Effects of environmental moisture, vegetation characteristics, and host abundance’, Ticks and Tick-borne Diseases, 11(1). Available at: https://doi.org/10.1016/j.ttbdis.2019.101271. Ginsberg, H.S. and Ewing, C.P. (1989) ‘Habitat Distribution of Ixodes dammini (Acari: Ixodidae) and Lyme Disease Spirochetes on Fire Island, New York’, Journal of Medical Entomology, 26(3), pp. 183–189. Available at: https://doi.org/10.1093/jmedent/26.3.183. Ginsberg, H.S. and Zhioua, E. (1996) ‘Nymphal survival and habitat distribution of Ixodes scapularis and Amblyomma americanum ticks (Acari: Ixodidae) on Fire Island, New York, USA’, Experimental and Applied Acarology, 20(9), pp. 533–544. Available at: https://doi.org/10.1007/BF00048285. Morgan, J.M. et al. (1994) ‘Predicting Ixodes scapularis Abundance on White-Tailed Deer Using Geographic Information Systems’, The American Journal of Tropical Medicine and Hygiene, 51(5), pp. 538–544. Available at: https://doi.org/10.4269/ajtmh.1994.51.538. Glass, G.E., Ganser, C. and Kessler, W.H. (2021) ‘Validating Species Distribution Models with Standardized Surveys for Ixodid Ticks in Mainland Florida’, Journal of Medical Entomology, 58(3), pp. 1345–1351. Available at: https://doi.org/10.1093/jme/tjaa282. Gourley, A.S. et al. (2018) ‘Role of white-tailed deer in geographic spread of the black-legged tick Ixodes scapularis: Analysis of a spatially nonlocal model’, Mathematical Biosciences & Engineering, 15(4), pp. 1033–1054. Available at: https://doi.org/10.3934/mbe.2018046. Guerra, M. et al. (2002) ‘Predicting the risk of Lyme disease: Habitat suitability for Ixodes scapularis in the north central United States’, Emerging Infectious Diseases, 8(3), pp. 289–297. Available at: https://doi.org/10.3201/eid0803.010166. Hahn, M.B. et al. (2016) ‘Modeling the Geographic Distribution of Ixodes scapularis and Ixodes pacificus (Acari: Ixodidae) in the Contiguous United States’, Journal of Medical Entomology, 53(5), pp. 1176–1191. Available at: https://doi.org/10.1093/jme/tjw076. Hamer, S.A. et al. (2010) ‘Invasion of the Lyme Disease vector Ixodes scapularis: Implications for Borrelia burgdorferi endemicity’, EcoHealth, 7(1), pp. 47–63. Available at: https://doi.org/10.1007/s10393-010-0287-0. Hamer, S.A. et al. (2014) ‘Increased diversity of zoonotic pathogens and Borrelia burgdorferi strains in established versus incipient Ixodes scapularis populations across the Midwestern United States’, Infection, Genetics and Evolution, 27, pp. 531–542. Available at: https://doi.org/10.1016/j.meegid.2014.06.003. 189 Hayes, L.E., Scott, J.A. and Stafford, K.C. (2015) ‘Influences of weather on Ixodes scapularis nymphal densities at long-term study sites in Connecticut’, Ticks and Tick-borne Diseases, 6(3), pp. 258–266. Available at: https://doi.org/10.1016/j.ttbdis.2015.01.006. Hijmans, R.J. et al. (2005) ‘Very high-resolution interpolated climate surfaces for global land areas’, International Journal of Climatology, 25(15), pp. 1965–1978. Available at: https://doi.org/10.1002/joc.1276. Hijmans, R.J. and van Etten, J. (2012) ‘raster: Geographic analysis and modeling with raster data. R package version 2.0-12.’ Illoldi-Rangel, P. et al. (2012) ‘Species distribution models and ecological suitability analysis for potential tick vectors of Lyme disease in Mexico’, Journal of Tropical Medicine, 2012. Available at: https://doi.org/10.1155/2012/959101. Jackson, J.O. and DeFoliart, G.R. (1970) ‘Ixodes scapularis Say in northern Wisconsin.’, Journal of medical entomology, 7(1), pp. 124–125. Available at: https://doi.org/10.1093/jmedent/7.1.124. James, A.M. et al. (2015) ‘The geographic distribution and ecological preferences of the American dog tick, Dermacentor variabilis (Say), in the U.S.A.’, Medical and Veterinary Entomology, 29(2), pp. 178–188. Available at: https://doi.org/10.1111/mve.12099. Johnson, T.L. et al. (2016) ‘Habitat suitability model for the distribution of Ixodes scapularis (acari: Ixodidae) in Minnesota’, Journal of Medical Entomology, 53(3), pp. 598–606. Available at: https://doi.org/10.1093/jme/tjw008. Keirans, J.E. and Litwak, T.R. (1989) ‘Pictorial Key to the Adults of Hard Ticks, Family Ixodidae (Ixodida: Ixodoidea), East of the Mississippi River’, Journal of medical entomology, 26(5), pp. 345–448. Available at: https://doi.org/10.1093/jmedent/26.5.435. Kessler, W.H., Ganser, C. and Glass, G.E. (2019) ‘Modeling the distribution of medically important tick species in Florida’, Insects, 10(7). Available at: https://doi.org/10.3390/insects10070190. Khatchikian, C.E. et al. (2015) ‘Recent and rapid population growth and range expansion of the Lyme disease tick vector, Ixodes scapularis, in North America’, Source: Evolution, 69(7), pp. 1678–1689. Available at: https://doi.org/10.1111/evo. Kitron, U. et al. (1992) ‘Spatial Analysis of the Distribution of Ixodes dammini (Acari: Ixodidae) on White-Tailed Deer in Ogle County, Illinois’, J. Med. Entomol, pp. 259–266. Available at: https://doi.org/10.1093/jmedent/29.2.259. Kopsco, H.L. et al. (2023) ‘Current and Future Habitat Suitability Models for Four Ticks of Medical Concern in Illinois, USA’, Insects, 14(3), p. 213. Available at: https://doi.org/10.3390/insects14030213. 190 Kopsco, H.L., Smith, R.L. and Halsey, S.J. (2022) ‘A Scoping Review of Species Distribution Modeling Methods for Tick Vectors’, Frontiers in Ecology and Evolution, 10. Available at: https://doi.org/10.3389/fevo.2022.893016. Kugeler, K.J. et al. (2015) ‘Geographic distribution and expansion of human Lyme Disease, United States’, Emerging Infectious Diseases, 21(8), pp. 1455–1457. Available at: https://doi.org/10.3201/eid2108.141878. Kugeler, K.J. et al. (2016) ‘Will Culling White-Tailed Deer Prevent Lyme Disease?’, Zoonoses and Public Health. Wiley-VCH Verlag, pp. 337–345. Available at: https://doi.org/10.1111/zph.12245. Lantos, P.M. et al. (2017) ‘Geographic expansion of Lyme disease in Michigan, 2000-2014’, Open Forum Infectious Diseases, 4(1), pp. 1–5. Available at: https://doi.org/10.1093/oid/ofw269. Larson, S.R. et al. (2022) ‘Ixodes scapularis density in US temperate forests shaped by deer, earthworms, and disparate factors at two scales’, Ecosphere, 13(2). Available at: https://doi.org/10.1002/ecs2.3932. Lee, X. et al. (2014) ‘Prevalence of Borrelia burgdorferi and Anaplasma phagocytophilum in Ixodes scapularis (Acari: Ixodidae) nymphs collected in managed red pine forests in Wisconsin.’, Journal of medical entomology, 51(3), pp. 694–701. Available at: https://doi.org/10.1603/ME13140. Leverett, F. and Taylor, F.B. (1915) ‘The Pleistocene of Indiana and Michigan and the history of the Great Lakes.’ Washington Government Printing Office. Available at: https://doi.org/10.3133/m53. Lieske, D.J. and Lloyd, V.K. (2018) ‘Combining public participatory surveillance and occupancy modelling to predict the distributional response of Ixodes scapularis to climate change’, Ticks and Tick-borne Diseases, 9(3), pp. 695–706. Available at: https://doi.org/10.1016/j.ttbdis.2018.01.018. Lindsay, L.R. et al. (1998) ‘Survival and Development of the Different Life Stages of Ixodes scapularis (Acari: Ixodidae) Held within Four Habitats on Long Point, Ontario’, Canada, J. Med. Entomol, pp. 189–199. Available at: https://doi.org/10.1093/jmedent/35.3.189. Lindsay, L.R. et al. (1999) ‘Microclimate and Habitat in Relation to Ixodes scapularis (Acari: Ixodidae) Populations on Long Point, Ontario, Canada’, Journal of Medical Entomology, 36(3), pp. 255–262. Available at: https://doi.org/10.1093/jmedent/36.3.255. Linske, M.A. et al. (2019) ‘Impacts of deciduous leaf litter and snow presence on nymphal Ixodes scapularis (Acari: Ixodidae) overwintering survival in coastal New England, USA’, Insects, 10(8). Available at: https://doi.org/10.3390/insects10080227. 191 Lippi, C.A., Gaff, H.D., White, A.L., St. John, H.K., et al. (2021) ‘Exploring the Niche of Rickettsia montanensis (Rickettsiales: Rickettsiaceae) Infection of the American Dog Tick (Acari: Ixodidae), Using Multiple Species Distribution Model Approaches’, Journal of Medical Entomology, 58(3), pp. 1083–1092. Available at: https://doi.org/10.1093/jme/tjaa263. Lippi, C.A., Gaff, H.D., White, A.L. and Ryan, S.J. (2021) ‘Scoping review of distribution models for selected Amblyomma ticks and Rickettsial group pathogens’, PeerJ, 9, pp. 1–19. Available at: https://doi.org/10.7717/peerj.10596. Lubelczyk, C.B. et al. (2004) ‘Habitat associations of Ixodes scapularis (Acari: Ixodidae) in Maine’, Environmental Entomology, 33(4), pp. 900–906. Available at: https://doi.org/10.1603/0046-225X-33.4.900. De Marco, P. and Nóbrega, C.C. (2018) ‘Evaluating collinearity effects on species distribution models: An approach based on virtual species simulation’, PLoS ONE, 13(9). Available at: https://doi.org/10.1371/journal.pone.0202403. McMaster, G.S. and Wilhelm, W.W. (1997) ‘Growing degree-days: one equation, two interpretations’, Agricultural and Forest Meteorology, 87(4), pp. 291–300. Available at: https://doi.org/10.1016/S0168-1923(97)00027-0. Naimi, B. and Araújo, M.B. (2016) ‘Sdm: A reproducible and extensible R platform for species distribution modelling’, Ecography, 39(4), pp. 368–375. Available at: https://doi.org/10.1111/ecog.01881. Needham, G R and Teel, P D (1991) ‘Off-Host Physiological Ecology of Ixodid Ticks’, Annual Review of Entomology, 36(1), pp. 659–681. Available at: https://doi.org/10.1146/annurev.en.36.010191.003303. Nelder, M.P. et al. (2016) ‘Human pathogens associated with the blacklegged tick Ixodes scapularis: a systematic review’, Parasites & Vectors, 9(1), p. 265. Available at: https://doi.org/10.1186/s13071-016-1529-y. Nuttall, P.A. (2022) ‘Climate change impacts on ticks and tick-borne infections’, Biologia, 77(6), pp. 1503–1512. Available at: https://doi.org/10.1007/s11756-021-00927-2. Ogden, N.H. et al. (2018) ‘Evidence for Geographic Variation in Life-Cycle Processes Affecting Phenology of the Lyme Disease Vector Ixodes scapularis (Acari: Ixodidae) in the United States’, Journal of Medical Entomology, 55(6), pp. 1386–1401. Available at: https://doi.org/10.1093/jme/tjy104. Ogden, N.H. and Lindsay, L.R. (2016) ‘Effects of Climate and Climate Change on Vectors and Vector-Borne Diseases: Ticks Are Different’, Trends in Parasitology. Elsevier Ltd, pp. 646–656. Available at: https://doi.org/10.1016/j.pt.2016.04.015. 192 Omernik, J.M. and Griffith, G.E. (2014) ‘Ecoregions of the Conterminous United States: Evolution of a Hierarchical Spatial Framework’, Environmental Management, 54(6), pp. 1249– 1266. Available at: https://doi.org/10.1007/s00267-014-0364-1. Ostfeld, R.S. et al. (1995) ‘Ecology of Lyme Disease: Habitat Associations of Ticks (Ixodes scapularis) In a Rural Landscape’, Ecological Applications, 5(2), pp. 353–361. Available at: https://doi.org/10.2307/1942027. Ostfeld, R.S. et al. (2006) ‘Climate, deer, rodents, and acorns as determinants of variation in Lyme-disease risk’, PLoS Biology, 4(6), pp. 1058–1068. Available at: https://doi.org/10.1371/journal.pbio.0040145. Ostfeld, R.S. and Brunner, J.L. (2015) ‘Climate change and Ixodes tick-borne diseases of humans’, Philosophical Transactions of the Royal Society B: Biological Sciences, 370(1665), pp. 1–11. Available at: https://doi.org/10.1098/rstb.2014.0051. Pfäffle, M. et al. (2013) ‘The ecology of tick-borne diseases’, International Journal for Parasitology, pp. 1059–1077. Available at: https://doi.org/10.1016/j.ijpara.2013.06.009. Poggio, L. et al. (2021) ‘SoilGrids 2.0: Producing soil information for the globe with quantified spatial uncertainty’, SOIL, 7(1), pp. 217–240. Available at: https://doi.org/10.5194/soil-7-217- 2021. Randolph, S.E. (2004) ‘Tick ecology: Processes and patterns behind the epidemiological risk posed by ixodid ticks as vectors’, Parasitology, 129(SUPPL.). Available at: https://doi.org/10.1017/S0031182004004925. Ripoche, M. et al. (2018) ‘Multi-scale clustering of Lyme disease risk at the expanding leading edge of the range of Ixodes scapularis in Canada’, International Journal of Environmental Research and Public Health, 15(4). Available at: https://doi.org/10.3390/ijerph15040603. Rocklöv, J. and Dubrow, R. (2020) ‘Climate change: an enduring challenge for vector-borne disease prevention and control’, Nature Immunology. Nature Research, pp. 479–483. Available at: https://doi.org/10.1038/s41590-020-0648-y. Rodgers, S.E., Zolnik, C.P. and Mather, T.N. (2007) ‘Duration of Exposure to Suboptimal Atmospheric Moisture Affects Nymphal Blacklegged Tick Survival’, Journal of Medical Entomology, 44(2), pp. 372–375. Available at: https://doi.org/10.1093/jmedent/44.2.372. Ruefenacht, B. et al. (2008) ‘Conterminous U.S. and Alaska Forest Type Mapping Using Forest Inventory and Analysis Data’, Photogrammetric Engineering & Remote Sensing, 74(11), pp. 1379–1388. Available at: https://doi.org/10.14358/PERS.74.11.1379. Rulison, E.L. et al. (2013) ‘Flagging Versus Dragging as Sampling Methods for Nymphal Ixodes scapularis (Acari: Ixodidae)’, Journal of Vector Ecology, 38(1), pp. 163–167. Available at: https://doi.org/10.1111/j.1948-7134.2013.12022.x. 193 Schulze, T.L., Jordan, R.A. and Hung, R.W. (1998) ‘Comparison of Ixodes scapularis (Acari: Ixodidae) Populations and their Habitats in Established and Emerging Lyme Disease Areas in New Jersey’, Journal of Medical Entomology, 35(1), pp. 64–70. Available at: https://doi.org/10.1093/jmedent/35.1.64. Schulze, T.L., Jordan, R.A. and Hung, R.W. (2001) ‘Effects of Selected Meteorological Factors on Diurnal Questing of Ixodes scapularis and Amblyomma americanum (Acari: Ixodidae)’, Journal of Medical Entomology, 38(2), pp. 318–324. Available at: https://doi.org/10.1603/0022- 2585-38.2.318. Scott, R.W. and Huff, F.A. (1996) ‘Impacts of the Great Lakes on regional climate conditions’, Journal of Great Lakes Research, 22(4), pp. 845–863. Available at: https://doi.org/10.1016/S0380-1330(96)71006-7. Slatculescu, A.M. et al. (2020) ‘Species distribution models for the eastern blacklegged tick, Ixodes scapularis, and the Lyme Disease pathogen, Borrelia burgdorferi, in Ontario, Canada’, PLoS ONE, 15(9 September), pp. 1–19. Available at: https://doi.org/10.1371/journal.pone.0238126. Somodi, I., Lepesi, N. and Botta-Dukát, Z. (2017) ‘Prevalence dependence in model goodness measures with special emphasis on true skill statistics’, Ecology and Evolution, 7(3), pp. 863– 872. Available at: https://doi.org/10.1002/ece3.2654. Sonenshine, D.E. (1979) ‘Insects of Virginia’, Research Division Bulletin. Virginia polytechnic institute and state university. Soucy, J.-P.P.R. et al. (2018) ‘High-Resolution Ecological Niche Modeling of Ixodes scapularis Ticks Based on Passive Surveillance Data at the Northern Frontier of Lyme Disease Emergence in North America’, Vector-Borne and Zoonotic Diseases, 18(5), pp. 235–242. Available at: https://doi.org/10.1089/vbz.2017.2234. Springer, Y.P. et al. (2015) ‘Modeling the present and future geographic distribution of the lone star tick, Amblyomma americanum (Ixodida: Ixodidae), in the continental United States’, American Journal of Tropical Medicine and Hygiene, 93(4), pp. 875–890. Available at: https://doi.org/10.4269/ajtmh.15-0330. Strand, M.R., Walker, E.D. and Merritt, R.W. (1992) ‘Field studies on Ixodes dammini in the Upper Peninsula of Michigan’, Vector Control Bulletin of North Central States, 1, pp. 11–18. Süss, J. et al. (2008) ‘What makes ticks tick? Climate change, ticks, and tick-borne diseases’, Journal of Travel Medicine, pp. 39–45. Available at: https://doi.org/10.1111/j.1708- 8305.2007.00176.x. 194 Swei, A. et al. (2020) ‘Patterns, Drivers, and Challenges of Vector-Borne Disease Emergence’, Vector-Borne and Zoonotic Diseases. Mary Ann Liebert Inc., pp. 159–170. Available at: https://doi.org/10.1089/vbz.2018.2432. Templer, P.H. et al. (2012) ‘Impact of a reduced winter snowpack on litter arthropod abundance and diversity in a northern hardwood forest ecosystem’, Biology and Fertility of Soils, 48(4), pp. 413–424. Available at: https://doi.org/10.1007/s00374-011-0636-3. Thornton, M.M. et al. (2022) ‘Daymet: Daily Surface Weather Data on a 1-km Grid for North America’, Version 4 R1., ORNL DAAC, Oak Ridge, Tennessee, USA. Available at: https://doi.org/10.3334/ORNLDAAC/2129. Thornton, P.E., Running, S.W. and White, M.A. (1997) ‘Generating surfaces of daily meteorological variables over large regions of complex terrain’, Journal of Hydrology, 190(3–4), pp. 214–251. Available at: https://doi.org/10.1016/S0022-1694(96)03128-9. U.S. Census Bureau (2022) ‘Total population, population density, sex ratio, and dependency ratios, by Local Health Departments, Michigan 2021’. U.S. Geological Survey (USGS) - Gap Analysis Project (GAP) (2018) ‘White-tailed Deer (Odocoileus virginianus) mWTDEx_CONUS_2001v1 Habitat Map: U.S. Geological Survey data release’. Vail, S.G. and Smith, G. (1998) ‘Air Temperature and Relative Humidity Effects on Behavioral Activity of Blacklegged Tick (Acari: Ixodidae) Nymphs in New Jersey’, Journal of Medical Entomology, 35(6), pp. 1025–1028. Available at: https://doi.org/10.1093/jmedent/35.6.1025. Valsson, S. and Bharat, A. (2011) ‘Impact of Air Temperature on Relative Humidity - A study’, Architecture - Time Space and People, (February), pp. 38–41. VanAcker, M.C. et al. (2019) ‘Enhancement of risk for Lyme disease by landscape connectivity, New York, New York, USA’, Emerging Infectious Diseases, 25(6), pp. 1136–1143. Available at: https://doi.org/10.3201/eid2506.181741. Vandyk, J.K. et al. (1996) ‘Survival of Ixodes scapularis (Acari: Ixodidae) Exposed to Cold’, Journal of Medical Entomology, 33(1), pp. 6–10. Available at: https://doi.org/10.1093/jmedent/33.1.6. Volk, M.R. et al. (2022) ‘Microclimate conditions alter Ixodes scapularis (Acari: Ixodidae) overwinter survival across climate gradients in Maine, United States’, Ticks and Tick-borne Diseases, 13(1). Available at: https://doi.org/10.1016/j.ttbdis.2021.101872. Vu, D.H., Muttaqi, K.M. and Agalgaonkar, A.P. (2015) ‘A variance inflation factor and backward elimination based robust regression model for forecasting monthly electricity demand using climatic variables’, Applied Energy, 140, pp. 385–394. Available at: https://doi.org/10.1016/j.apenergy.2014.12.011. 195 Walker, E.D. et al. (1994) ‘Prevalence of Borrelia burgdorferi in Host-Seeking Ticks (Acari: Ixodidae) from a Lyme Disease Endemic Area in Northern Michigan’, Journal of Medical Entomology, 31(4), pp. 524–528. Available at: http://www.ingentaconnect.com/content/esa/jme/1994/00000031/00000004/art00003. Walker, E.D. et al. (1998) ‘Geographic Distribution of Ticks (Acari: Ixodidae) in Michigan, with Emphasis on Ixodes scapularis and Borrelia burgdorferi’, Journal of Medical Entomology, 35(5), pp. 872–882. Available at: https://doi.org/10.1093/jmedent/35.5.872. Wallace, D. et al. (2019) ‘Effect of Rising Temperature on Lyme Disease: Ixodes scapularis Population Dynamics and Borrelia burgdorferi Transmission and Prevalence’, Canadian Journal of Infectious Diseases and Medical Microbiology, 2019, pp. 22–24. Available at: https://doi.org/10.1155/2019/9817930. Wickham, J. et al. (2021) ‘Thematic accuracy assessment of the NLCD 2016 land cover for the conterminous United States’, Remote Sensing of Environment, 257. Available at: https://doi.org/10.1016/j.rse.2021.112357. Witmer, G., Gary, W. and Decalesta, D.S. (1991) ‘The need and difficulty of bringing Pennsylvania deer herd under control’, in Eastern Wildlife Damage Control Conference, pp. 130–137. Wolf, M.J., Watkins, H.R. and Schwan, W.R. (2020) ‘Ixodes scapularis: Vector to an Increasing Diversity of Human Pathogens in the Upper Midwest.’, WMJ: official publication of the State Medical Society of Wisconsin, 119(1), pp. 16–21. Available at: http://www.ncbi.nlm.nih.gov/pubmed/32348066. Yang, L. et al. (2018) ‘A new generation of the United States National Land Cover Database: Requirements, research priorities, design, and implementation strategies’, ISPRS Journal of Photogrammetry and Remote Sensing, 146, pp. 108–123. Available at: https://doi.org/10.1016/j.isprsjprs.2018.09.006. 196 APPENDIX Figure 5.9. The sampling sites of Michigan from 2017 to 2021 for drag sampling Ixodes scapularis. 197 Figure 5.10. County establishment data of Ixodes scapularis from 2017 to 2021 according to CDC criteria (Dennis et al. 1998; Eisen et al. 2016). Established criteria correspond to counties that had greater than 5 I. scapularis of any life stage detected or 2 life stages of I. scapularis detected over one calendar year. Reported criteria correspond to the detection of I. scapularis but at levels below that needed to be considered ‘established’. 198 Table 5.7. The Ixodes scapularis presence categories that SDMs were based on for the whole state of Michigan. Category Definition Lenient Any site that at least one I. scapularis was detected Moderate Any sites that had established populations of I. scapularis (i.e., sites that had greater than 5 I. scapularis of any life stage or 2 life stages of I. scapularis detected over one calendar year). Stringent Any sites that metboth establishment criteria I. scapularis populations present (i.e. sites that had greater than 5 I. scapularis of any life stage and 2 life stages of I. scapularis detected over one calendar year). 199 Figure 5.11. The GLM and MaxEnt models for the presence of Ixodes scapularis for the entire state of Michigan from the lenient category (A-B) to moderate category (B-C) and the stringent category (E – F) of I. scapularis presence/established levels. 200 CHAPTER 6: CONCLUSION Ixodes scapularis is rapidly spreading throughout the northeastern and north central regions of the US. With the increase in the geographic distribution of I. scapularis there also has been an increase in human diseases that are vectored by I. scapularis such as Lyme disease, human granulocytic anaplasmosis, babesiosis and Powassan virus encephalitis (Bacon, Kugeler and Mead, 2008; Kugeler et al., 2015; Adams et al., 2017; Schwartz et al., 2017). Improving our knowledge about the patterns and mechanisms of spread of I. scapularis and its associated pathogens will enable us to develop prevention and mitigation strategies. Lyme disease is the most common vector-borne disease in the US, and much of the knowledge that is derived from ecological studies that have been conducted on the Lyme disease bacterium can serve to inform the ecology of other tick-borne pathogens vectored by I. scapularis. Here I wanted to investigate further the disease ecology of Anaplasma phagocytophilum, the causative agent of human granulocytic anaplasmosis, the second most common vector borne disease in the US (Adams et al., 2017). Anaplasma phagocytophilum, a gram-negative bacterium, infects white blood cells in mammals. The disease symptoms are very similar to Lyme disease (Dumler et al., 2005; Bakken and Dumler, 2015; Sanchez et al., 2016; Baker et al., 2020), so often in endemic regions of Lyme disease such as the Northeast and the Midwest human granulocytic anaplasmosis may be mistaken for Lyme disease. As the treatment for anaplasmosis is similar to that of Lyme disease, anaplasmosis may be underreported. Ecology of A. phagocytophilum at an I. scapularis endemic site in the Upper Midwest The ecology of A. phagocytophilum has been studied at a site in upstate New York where the reservoir competence of small to medium sized mammals were estimated (Keesing et al., 2012, 2014). We wanted to conduct a similar study in the Upper Midwest because there were not any 201 studies available that looked at several host species over the same time period. We quantified the infection prevalence and the reservoir competence of small to medium sized reservoir hosts; characterized host features that are important in becoming infected with A. phagocytophilum, including features that would influence larval and nymphal I. scapularis counts on hosts; and characterized the phenology of infection in hosts in relationship to parasitizing nymphs and larvae. We also wanted to compare detection of A. phagocytophilum of hosts using two types of tissues (blood vs ear biopsy tissues). Overall, our results support that found in other studies conducted in the Northeast and limited data in the Midwest when considering larval and nymphal burdens among species. We focused our comparison regarding A. phagocytophilum-ha infection among hosts with that of Keesing et al. (2014) because they trapped multiple species of host at one site simultaneously as we had, and because they strain-typed their positives. Keesing et al. (2014) found that the two species with the highest realized reservoir competence were the white-footed mouse (Peromyscus leucopus) and the eastern chipmunk (Tamias striatus). The realized reservoir competence of eastern chipmunks and white-footed mice for A. phagocytophilum at Ft. McCoy from 2010-2012 were much higher than that found at Keesing et al (2014). In fact, the eastern chipmunk had the highest infection prevalence for A. phagocytophilum. Because white-footed mice fed the most larvae compared to other species, however, they may contribute the most to infecting larvae with A. phagocytophilum. A limitation of this study is that we did not have large enough sample sizes of most mammals to compare the realized reservoir competence, because we were not able to collect enough engorged larvae from other hosts species to have confidence in estimating infectivity; therefore, we only compared the white-footed mouse and eastern chipmunk. If, however, the relative proportion of larvae fed by host species captured in our study is 202 representative of the host community at Fort McCoy, it still appears that white-footed mice might contribute the most to infecting larvae with A. phagocytophilum, and therefore its enzootic cycle. It has been recognized that the relative timing of questing activity, i.e., phenology, of nymphal and larval I. scapularis can be important for the enzootic maintenance of pathogens that are not vertically transmitted (Ogden et al. 2007). For pathogens that have short-lived infections in the host, where the reservoir hosts may be short-lived, and/or where there might be high turnover rate in the reservoir host population, greater overlap (i.e., “synchrony”) is needed for maintaining the enzootic cycle. Laboratory studies have shown that although mice may harbor A. phagocytophilum infections for twelve weeks (Levin et al. 2004), the highest rates of transmission to larvae occur in the first 1-3 weeks post-infection. This is in comparison with B. burgdorferi, which can be transmit by mice at relatively high rates for life (Donahue, Piesman and Spielman, 1987). At Ft. McCoy, although there are questing larvae throughout the summer, the phenology of infected on-host larvae peaks in the first half of summer, which reflects the phenologies of questing infected nymphs and infected mice, which also peak in the first half of summer, as one would expect. The seeming coincident timing of the peak in questing infected nymphs and infected on- host larvae also suggests that co-infection (i.e., non-systemic) transmission might occur. This is of interest because in the Northeast, although the nymphal peak occurs at the same time as in the Midwest, the larval peak occurs in late summer. Our data thus suggest that all else equal, the proportion of larvae that would become infected by feeding on white-footed mice would be much less in the Northeast compared to the Midwest, due to lower rates of systemic and non-systemic transmission. Phenology, among other factors, might contribute to the higher A. phagocytophilum- ha infection prevalence observed among questing nymphs, chipmunks, and mice at Ft. McCoy compared to southeastern New York (Keesing et al. 2014). 203 We looked factors that affect the probability that a host could become infected with A. phagocytophilum. Of interest, we found that individuals that were infected with B. burgdorferi had relatively a greater chance of being infected with A. phagocytophilum. At an I. scapularis endemic site like Fort McCoy, Wisconsin the chances of becoming infected with both B. burgdorferi and A. phagocytophilum would be greater and many hosts species can be coinfected. In other regions where I. scapularis is endemic, coinfection of B. burgdorferi and A. phagocytophilum in hosts is also quite common (Horowitz et al., 2013; Diuk-Wasser, Vannier and Krause, 2016; Westwood, Peters and Rooney, 2020b; Lehane et al., 2021) because of the similar ecologies shared by both pathogens. Finally, for simply determining the infection prevalence of hosts for A. phagocytophilum, our study suggests that once can sample either blood (which is much more commonly performed) or ear tissue biopsies. As A. phagocytophilum is an intracellular pathogen that infects white blood cells (Chen et al., 1994), blood is assayed in most cases. There was no significant difference between infection prevalence estimated from blood, biopsy and and/or both types of tissues simultaneously when we tested white-footed mice for whom we had collected both types of samples simultaneously. Ear biopsies have been used in some studies to look at the infection status of A. phagocytophilum (Baráková et al., 2014; Rosso et al., 2017) in Europe. A study conducted on canine skin lesions showed that even when dogs seroconverted and no longer showed signs of disease, A. phagocytophilum was detected in skin biopsies (Berzina et al., 2014), which was then hypothesized to be due to A. phagocytophilum being persistent in skin of hosts (Granquist, Aleksandersen, et al., 2010). Conducting a xenodiagnostic study using mice would help characterize the relationship of the timing of detection of A. phagocytophilum [DNA] in blood, biopsy, and transmission to feeding larvae. If skin biopsies could be as used a reliable indicator 204 of infection of A. phagocytophilum, it would be welcome knowledge, as it is less invasive and is easier to store and process. The case incidence of A. phagocytophilum has increased over time with the spread of I. scapularis (Biggs, Behravesh, K. K. Bradley, et al., 2016; Adams et al., 2017; Schwartz et al., 2017). Thus, conducting ecological studies across the geographic range of I. scapularis, where climate and wildlife host communities may vary, will help us further understand how A. phagocytophilum and its various strains are maintained enzootically. We can apply the knowledge gained to predict the level of public health risk of disease as well to control and prevent the further spread of A. phagocytophilum. Species Distribution modeling of I. scapularis in Michigan Michigan is at one of the leading edges of I. scapularis invasion. From the time ticks were first discovered in the early 1990s in the Upper Peninsula (Strand, Walker and Merritt, 1992) and in the 2000s in the Lower Peninsula (Erik S. Foster, 2004), blacklegged ticks have spread gradually throughout the state (Dennis et al., 1998; Eisen, Eisen and Beard, 2016; Lantos et al., 2017; Fleshman et al., 2021). Because of this gradual spread of I. scapularis, we wanted to develop suitable habitat models to identify the distribution of suitable habitats for the establishment of I. scapularis to help inform future disease risk. We also wanted to look at the climatic and habitat predictors that are potentially important in the spread and establishment of I. scapularis. In the literature there are several different types of species distribution modeling (SDM) techniques. In our study, we assessed two techniques- one which is based on presence-absence data (a regression- based generalized linear model (GLM) method) and the other, which is based on presence only data (machine learning-based Maximum Entropy (MaxEnt) model). We used active I. scapularis surveillance data collected from 2017 – 2021 from sites 205 located throughout Michigan. We used 16-19 environmental variables in the final models, where we modeled the Lower and Upper Peninsulas separately given the vast environmental and ecological differences. Our models had a relatively high area under the curve (AUC) of the receiver operator curve (ROC) values, which ranged from about 0.7 – 0.9, indicating that the model could categorize the presence or presence-absence points better than by random chance. When considering the climatic and habitat predictors that were important for the model, the overall climatic factors that were related to temperature and humidity seem to be most important. Both temperature and humidity are two climatic factors that affect the survival of I. scapularis because I. scapularis is very susceptible to desiccation (Vail and Smith, 1998; Berger et al., 2014; Elias et al., 2021). The habitat features that were important were factors that were related to the presence of white-tailed deer, presence of deciduous woods, elevation, soil moisture, amount of leaf litter. A caveat is that because our sampling protocol was designed to find I. scapularis ticks, we targeted habitats that were believed to be suitable habitats for I. scapularis; this may have constrained the habitat variables that came out as important variables in our model. We found general agreement between the modeling techniques for predicting distributions of suitable habitats for both the Upper and Lower Peninsulas. Suitable habitats in the Lower Peninsula were more concentrated in the southern portion of the state and within pockets along the west and east coast of the state, while in the Upper Peninsula suitable habitats were concentrated in the southern central region bordering Wisconsin and along Lake Superior also in the central region. A recent study conducted using passive and actively collected data collated at the county level from throughout the Northeast and the Midwest showed most of Michigan to have suitable habitat for the establishment of I. scapularis (Burtis et al., 2022). It was different from our study 206 since we found regions in the northern central region of the Lower Peninsula and in the eastern portion of the Upper Peninsula to be unsuitable for I. scapularis. Because our model was based on the current I. scapularis data from Michigan and because I. scapularis is still emerging across Michigan, our model likely is based on data that includes “false negatives”. Given our understanding of I. scapularis ecology and the range of sites in which it already is established throughout the eastern U.S., we hypothesize that I. scapularis eventually will become widespread throughout the state. Therefore, some of the sites and regions that our models predicted to be unsuitable for ticks might be truly unsuitable or may be areas I. scapularis has not gotten to yet. A way to test this hypothesis directly would be set up tick gardens (i.e, semi-natural arenas) in those regions that the model predicted to be unsuitable and release I. scapularis to measure their survival and development rates. Another would be to continue with periodic active surveillance focused on those regions to detect the presence and establishment of I. scapularis. We could also develop species distribution models based on data from Wisconsin, where I. scapularis has time to spread throughout the state. Those models could be projected onto Michigan as Foster (2004) had done with Guerra et al. (2002), leading to the discovery of I. scapularis in southwest Michigan. This may provide us with a better understanding if some regions in the Upper and Lower Peninsula are indeed unsuitable. Using Wisconsin as the base model is beneficial because the climatic and habitat variables are similar to Michigan. To further improve our model, we could use other techniques to model the distribution of I. scapularis and focus on developing ensemble models, which would give us a better understanding about how each of the SDMs work. Since we have I. scapularis abundance data and density data developing models based on these would also give us a better understanding of patterns of invasion of I. scapularis. Developing SDMs is one way to predict species distributions that is relatively inexpensive 207 and less time consuming if active surveillance, especially in larger regions, cannot be done. Using SDM’s also can help guide surveillance activities to regions that have suitable habitats such that one can try to detect invading I. scapularis when they are still at low density. These SDMs thus can guide planning prevention and control strategies especially in areas that I. scapularis have not been established. As with any type of modeling effort, however, researchers need to be aware of the limitations imposed by the input data and the model assumptions as well as how the models work. Ticks and tickborne disease have been an important topic in modern times as we observe an increase in several tickborne disease cases like Lyme disease, anaplasmosis, babesiosis, Powassan virus encephalitis, ehrlichiosis, Rocky Mountain spotted fever, as well as the tick- associated red meat allergy. With climate change, habitat fragmentation and modification, tick distributions will shift and/or expand into regions where they have not been observed in recent time. Developing SDMs and trying to gain more insight into how tickborne pathogens are maintained in nature by the wildlife and vector communities will help in the fight for prevention and control of ticks and tickborne diseases. 208 BIBLIOGRAPHY Adams, D.A. et al. (2017) ‘Summary of Notifiable Infectious Diseases and Conditions — United States, 2015’, MMWR. Morbidity and Mortality Weekly Report, 64(53), pp. 1–143. Available at: https://doi.org/10.15585/mmwr.mm6453a1. Bacon, R.M., Kugeler, K.J. and Mead, P.S. (2008) ‘Surveillance for Lyme disease-United States, 1992-2006., MMWR. CDC Surveillance summaries : Morbidity and mortality weekly report’. Available at: https://www.cdc.gov/mmwr/preview/mmwrhtml/ss5710a1.htm (Accessed: 4 July 2023). Baker, A. et al. (2020) ‘Increasing Incidence of Anaplasmosis in the United States, 2012 through 2016’, Vector-Borne and Zoonotic Diseases, 20(11), pp. 855–859. Available at: https://doi.org/10.1089/vbz.2019.2598. Bakken, J.S. and Dumler, J.S. (2015) ‘Human Granulocytic Anaplasmosis’, Infect Dis Clin of North Am., 29(1), pp. 341–355. Available at: https://doi.org/10.1016/j.idc.2015.02.007. Baráková, I. et al. (2014) ‘Genetic and ecologic variability among Anaplasma phagocytophilum strains, Northern Italy’, Emerging Infectious Diseases, 20(6), pp. 1082–1085. Available at: https://doi.org/10.3201/eid2006.131023. Berger, K.A. et al. (2014) ‘Relative humidity and activity patterns of Ixodes scapularis (Acari: Ixodidae)’, Populations and Community Ecology, 51(4), pp. 769–776. Available at: https://doi.org/https://doi.org/10.1603/ME13186. Berzina, I. et al. (2014) ‘Anaplasma phagocytophilum DNA amplified from lesional skin of seropositive dogs’, Ticks and Tick-borne Diseases, 5(3), pp. 329–335. Available at: https://doi.org/10.1016/j.ttbdis.2013.12.010. Biggs, H.M. et al. (2016) ‘Diagnosis and Management of Tickborne Rickettsial Diseases: Rocky Mountain Spotted Fever and Other Spotted Fever Group Rickettsioses, Ehrlichioses, and Anaplasmosis - United States.’, MMWR. Recommendations and reports : Morbidity and mortality weekly report. Recommendations and reports, 65(2), pp. 1–44. Available at: https://doi.org/10.15585/mmwr.rr6502a1. Bouchard, C. et al. (2011) ‘Associations between Ixodes scapularis ticks and small mammal hosts in a newly endemic zone in southeastern Canada: Implications for Borrelia burgdorferi transmission’, Ticks and Tick-borne Diseases, 2(4), pp. 183–190. Available at: https://doi.org/10.1016/j.ttbdis.2011.03.005. Brunner, J.L. and Ostfeld, R.S. (2008) ‘Multiple causes of variable tick burdens on small- mammal hosts’, Ecology, 89(8), pp. 2259–2272. Available at: https://doi.org/10.1890/07-0665.1. Burtis, J.C. et al. (2022) ‘Predicting distributions of blacklegged ticks (Ixodes scapularis), Lyme disease spirochetes (Borrelia burgdorferi sensu stricto) and human Lyme disease cases in the 209 eastern United States’, Ticks and Tick-borne Diseases, 13(5). Available at: https://doi.org/10.1016/j.ttbdis.2022.102000. Chen, S.M. et al. (1994) ‘Identification of a granulocytotropic Ehrlichia species as the etiologic agent of human disease.’, Journal of Clinical Microbiology, 32(3), pp. 589–595. Available at: https://doi.org/10.1128/jcm.32.3.589-595.1994. Dennis, D.T. et al. (1998) ‘Reported Distribution of Ixodes scapularis and Ixodes pacificus (Acari: Ixodidae) in the United States’, Journal of Medical Entomology, 35(5), pp. 629–638. Available at: https://doi.org/10.1093/jmedent/35.5.629. Diuk-Wasser, M.A., Vannier, E. and Krause, P.J. (2016) ‘Coinfection by Ixodes Tick-Borne Pathogens: Ecological, Epidemiological, and Clinical Consequences’, Trends in Parasitology, 32(1), pp. 30–42. Available at: https://doi.org/10.1016/j.pt.2015.09.008. Dumler, J.S. et al. (2005) ‘Human granulocytic anaplasmosis and Anaplasma phagocytophilum’, Emerging Infectious Diseases, 11(12), pp. 1828–1834. Available at: https://doi.org/10.3201/eid1112.050898. Eisen, R.J., Eisen, L. and Beard, C.B. (2016) ‘County-scale distribution of Ixodes scapularis and Ixodes pacificus (Acari: Ixodidae) in the continental United States’, Journal of Medical Entomology, 53(2), pp. 349–386. Available at: https://doi.org/10.1093/jme/tjv237. Elias, S.P. et al. (2021) ‘A generalized additive model correlating blacklegged ticks with white- tailed deer density, temperature, and humidity in Maine, USA, 1990-2013’, Journal of Medical Entomology, 58(1), pp. 125–138. Available at: https://doi.org/10.1093/jme/tjaa180. Fleshman, A.C. et al. (2021) ‘Reported County-Level Distribution of Lyme Disease Spirochetes, Borrelia burgdorferi sensu stricto and Borrelia mayonii (Spirochaetales: Spirochaetaceae), in Host-Seeking Ixodes scapularis and Ixodes pacificus Ticks (Acari: Ixodidae) in the Contiguous United States’, Journal of Medical Entomology. Edited by M. Diuk-Wasser, 58(3), pp. 1219– 1233. Available at: https://doi.org/10.1093/jme/tjaa283. Foster, E.S. (2004)’ Ixodes scapularis (Acari: Ixodidae) and Borrelia burgdorferi in southwest Michigan: population ecology and verification of a geographic risk model’. Michigan State University. Granquist, E.G. et al. (2010) ‘A morphological and molecular study of Anaplasma phagocytophilum transmission events at the time of Ixodes ricinus tick bite.’, Acta veterinaria Scandinavica, 52(1), p. 43. Available at: https://doi.org/10.1186/1751-0147-52-43. Horowitz, H.W. et al. (2013) ‘Lyme Disease and human granulocytic anaplasmosis coinfection: Impact of case definition on coinfection rates and illness severity’, Clinical Infectious Diseases, 56(1), pp. 93–99. Available at: https://doi.org/10.1093/cid/cis852. Keesing, F. et al. (2012) ‘Reservoir Competence of Vertebrate Hosts for Anaplasma 210 phagocytophilum’, Emerging Infectious Diseases, 18(12), pp. 10–13. Available at: https://doi.org/10.3201/eid1812.120919. Keesing, F. et al. (2014) ‘Prevalence of human-Active and variant 1 strains of the tick-borne pathogen Anaplasma phagocytophilum in hosts and forests of Eastern North America’, American Journal of Tropical Medicine and Hygiene, 91(2), pp. 302–309. Available at: https://doi.org/10.4269/ajtmh.13-0525. Kugeler, K.J. et al. (2015) ‘Geographic distribution and expansion of human Lyme Disease, United States’, Emerging Infectious Diseases, 21(8), pp. 1455–1457. Available at: https://doi.org/10.3201/eid2108.141878. Lantos, P.M. et al. (2017) ‘Geographic expansion of Lyme Disease in Michigan, 2000-2014’, Open Forum Infectious Diseases, 4(1), pp. 1–5. Available at: https://doi.org/10.1093/oid/ofw269. Lehane, A. et al. (2021) ‘Prevalence of single and coinfections of human pathogens in Ixodes ticks from five geographical regions in the United States, 2013–2019’, Ticks and Tick-borne Diseases, 12(2), p. 101637. Available at: https://doi.org/10.1016/j.ttbdis.2020.101637. Rosso, F. et al. (2017) ‘Prevalence and genetic variability of Anaplasma phagocytophilum in wild rodents from the Italian alps’, Parasites & Vectors, 10(1), p. 293. Available at: https://doi.org/10.1186/s13071-017-2221-6. Sanchez, E. et al. (2016) ‘Diagnosis, Treatment, and Prevention of Lyme Disease, Human Granulocytic Anaplasmosis, and Babesiosis’, Jama, 315(16), p. 1767. Available at: https://doi.org/10.1001/jama.2016.2884. Schwartz, A.M. et al. (2017) ‘Surveillance for Lyme Disease — United States, 2008–2015’, MMWR. Surveillance Summaries, 66(22), pp. 1–12. Available at: https://doi.org/10.15585/mmwr.ss6622a1. Strand, M.R., Walker, E.D. and Merritt, R.W. (1992) ‘Field studies on Ixodes dammini in the Upper Peninsula of Michigan’, Vector Control Bulletin of North Central States, 1, pp. 11–18. Vail, S.G. and Smith, G. (1998) ‘Air Temperature and Relative Humidity Effects on Behavioral Activity of Blacklegged Tick (Acari: Ixodidae) Nymphs in New Jersey’, Journal of Medical Entomology, 35(6), pp. 1025–1028. Available at: https://doi.org/10.1093/jmedent/35.6.1025. Westwood, M.L., Peters, J.L. and Rooney, T.P. (2020) ‘Prevalence and Coinfection of Three Tick-Borne Pathogens in Questing Adult Blacklegged Ticks Ixodes scapularis (Vilas County, Wisconsin)’, Vector-Borne and Zoonotic Diseases, 20(8), pp. 633–635. Available at: https://doi.org/10.1089/vbz.2020.2619. 211