GENETIC MONITORING OF CUCURBIT DOWNY MILDEW IN MICHIGAN 

By 

Julian Camilo Bello Rodriguez 

 

 

 

 

 

 

A DISSERTATION 

Submitted to 

 

2020 

 

 

 

 

 

Michigan State University 

in partial fulfillment for the requirements 

for the degree of 

Plant Pathology – Doctor of Philosophy 

ABSTRACT 

GENETIC MONITORING OF CUCURBIT DOWNY MILDEW IN MICHIGAN 

By 

Julian Camilo Bello Rodriguez 

Cucurbit downy mildew (CDM) caused by the oomycete obligate pathogen, 

Pseudoperonospora cubensis, incites foliar blighting of several cucurbit genera 

worldwide. In 2004, the pathogen re-emerged in the U.S. infecting historically resistant 

cucumber cultivars and requiring the adoption of an intensive fungicide program. Due to 

an influx of aerially dispersed sporangia from overwinter sources, CDM occurs annually 

in cucumber growing regions of northern U.S. The genetic monitoring of incoming P. 

cubensis populations is essential for growers to make informed decisions regarding 

CDM management strategies. However, the scale and resolution of genetic studies of 

downy mildews (Peronosporaceae) remains limited due to the logistical constraints 

involved in the genotyping of these species (e.g. obtaining DNA of sufficient quantity 

and quality). To gain an evolutionary and ecological perspective of P. cubensis, we 

describe a targeted enrichment (TE) protocol able to genotype environmental samples 

of Pseudoperonospora spp. using less than 50 ng of DNA for library preparation. Using 

the TE protocol, we were able to enrich 736 target genes across 101 samples and 

identified 2,978 high quality SNP variants. This SNPs resolved the population structure 

of P. cubensis in Michigan and detected significant (AMOVA, P=0.01) genetic 

differentiation among the P. cubensis populations from squash (clade I) and cucumber 

(clade II). No evidence of location-based differentiation was detected within the P. 

cubensis (clade II) subpopulation of Michigan.  

Timely alerts of an influx of airborne inoculum of two distinct host-adapted clades 

of P. cubensis can assist Michigan growers in assessing the need to initiate fungicide 

sprays. However, the inability to distinguish between the morphologically identical 

sporangia of P. humuli and P. cubensis has been a significant shortcoming. Using spore 

traps and qPCR assays, an improved methodology for the aerial monitoring of each 

Pseudoperonospora taxa was identified. A highly specific qPCR assay differentiated 

Pseudoperonospora humuli, the causal agent of downy mildew on hop, and the two 

host-adapted clades of P. cubensis (clade I and II) on spore trap samples. After two 

years of monitoring, P. cubensis clade II DNA was detected in spore trap samples >2 

days before CDM symptoms were first observed in commercial cucumber fields 

(August), while P. humuli DNA was only detected early in the growing season (May and 

June). P. cubensis clade I DNA was not detected in air samples before or after the 

disease onset in cucumber fields. Additionally, the probability for P. cubensis detection 

in Burkard spore trap samples was higher compared to impaction spore trap samples 

with approximately the same number of sporangia, suggesting that the efficiency of 

recovery of sporangia by Burkard spore traps exceeds the recovery of impaction spore 

traps. The methodology described in this study to monitor the airborne concentrations of 

Pseudoperonospora spp. sporangia could be used as part of a CDM risk advisory 

system to time fungicide applications that protect cucurbit crops in Michigan. 

 

 

 

 

 

 

ACKNOWLEDGMENTS 

I would like to thank my co-advisor, Dr. Mary Hausbeck, for taking me as her 

student and for the work-related support she provided to me over the years. It was an 

honor to be part of her program and I'm thankful for your patience with my 

stubbornness. I would also like to thank my co-advisor Dr. Monique Sakalidis, for her 

guidance and questions, they challenged me and encourage me to continue in the 

graduate program. In addition, Dr. Andy Jarosz and Dr. Timothy Miles, for their 

invaluable advice as members of my committee.  

 

I could not have achieved my degree without the assistance of the technical staff 

and graduate students of the Hausbeck and Sakalidis Labs. Special thanks to Doug 

Higgins, my program fellow, for the countless hours of discussions after 5 p.m. Your 

insights, critics, and suggestions were essential to complete my objectives.  

 

I thank my parents for investing in my undergraduate program and for being 

there for me under any circumstances. I love you both. I also thank my sisters and 

brother; I made it to this program because you are always challenging me. And last, but 

not least, I would like to thank my good friends, Marcela Tabares, Oscar Benitez, and 

Clarissa Strieder, for their emotional support during the hardest times of my time far 

from home, you guys help me build the best version of myself. 

 

 

iv 

 

 

TABLE OF CONTENTS 

 
LIST OF TABLES ........................................................................................................... vii 
 
LIST OF FIGURES .......................................................................................................... ix 
 
CHAPTER I: LITERATURE REVIEW - PSEUDOPERONOSPORA CUBENSIS THE 
CAUSAL AGENT OF CUCURBIT DOWNY MILDEW ..................................................... 1 
INTRODUCTION .......................................................................................................... 2 
TAXONOMY AND DISEASE CYCLE ........................................................................... 5 
SURVIVAL AND DISPERSAL OF P. CUBENSIS ........................................................ 8 
GENETIC STRUCTURE ............................................................................................. 10 
PATHOGENIC VARIATION ........................................................................................ 12 
FUNGICIDE RESISTANCE ........................................................................................ 13 
BREEDING EFFORTS ............................................................................................... 18 
EARLY DETECTION OF P. CUBENSIS ..................................................................... 21 
LITERATURE CITED ..................................................................................................... 25 
 
CHAPTER II: DETECTION OF AIRBORNE SPORANGIA OF 
PSEUDOPERONOSPORA CUBENSIS AND P. HUMULI IN MICHIGAN USING 
BURKARD SPORE TRAPS COUPLED TO QPCR ....................................................... 39 
ABSTRACT ................................................................................................................. 40 
INTRODUCTION ........................................................................................................ 41 
MATERIALS AND METHODS .................................................................................... 43 
RESULTS ................................................................................................................... 48 
DISCUSSION .............................................................................................................. 54 
APPENDIX ..................................................................................................................... 60 
LITERATURE CITED ..................................................................................................... 71 
 
CHAPTER III: OPTIMIZING SPORE TRAPS AND QUANTITATIVE PCR ASSAYS 
FOR THE MONITORING OF CUCURBIT DOWNY MILDEW. ...................................... 77 
ABSTRACT ................................................................................................................. 78 
INTRODUCTION ........................................................................................................ 79 
MATERIALS AND METHODS .................................................................................... 82 
RESULTS ................................................................................................................... 86 
DISCUSSION .............................................................................................................. 93 
APPENDIX ................................................................................................................... 101 
LITERATURE CITED ................................................................................................... 127 
 
 
 

 

v 

 

 

CHAPTER IV: GENOTYPING OF THE OBLIGATE PLANT PATHOGENS 
PSEUDOPERONOSPORA CUBENSIS AND P. HUMULI USING TARGET 
ENRICHMENT SEQUENCING .................................................................................... 133 
ABSTRACT ............................................................................................................... 134 
INTRODUCTION ...................................................................................................... 135 
MATERIALS AND METHODS .................................................................................. 138 
RESULTS ................................................................................................................. 148 
DISCUSSION ............................................................................................................ 156 
APPENDIX ................................................................................................................... 164 
LITERATURE CITED ................................................................................................... 194 
 
CHAPTER V ................................................................................................................. 202 
CONCLUSIONS ........................................................................................................ 203 

 

 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

vi 

 

 

LIST OF TABLES 

 
Table 1-1. List of fungicides registered for the control of cucumber downy mildew in the 
U.S. ................................................................................................................................. 16 
 
Table 2-1. Primers and locked nucleic acids (LNA) probes for the qPCR assay 
differentiating Pseudoperonospora cubensis and P. humuli using the 105 SNP in the 
mitochondrial Cox2 gene. ............................................................................................... 61 
 
Table 2-2. Threshold cycle (Cq) values of the qPCR assays using LNA probes and 
varying concentrations of genomic DNA. ....................................................................... 62 
 
Table 3-1. qPCR assays designed for the differentiation of Pseudoperonospora humuli 
and Pseudoperonospora cubensis clade 1 and 2 targeting mitochondrial DNA regions.
 ...................................................................................................................................... 101 
 
Table 3-2. Threshold cycle (Cq) values of the qPCR protocol B using varying 
concentrations of genomic DNA from Pseudoperonospora cubensis clade 1 and 2 and 
P. humuli....................................................................................................................... 104 
 
Table 3-3. Percentage of qPCR positive samples collected by Burkard and impaction 
spore traps ................................................................................................................... 106 
 
Table 3-4. Linear regression analysis of Cq values as a function of the number 
Pseudoperonospora spp. sporangia ............................................................................ 107 
 
Table 3-5. b estimates of logistical models developed to predict the probability of 
Pseudoperonospora cubensis detection in Burkard and impaction spore trap samples
 ...................................................................................................................................... 108 
 
Table 3-6. Detection of Pseudoperonospora cubensis using Burkard spore traps 
coupled with qPCR and detection of symptoms ........................................................... 109 
 
Table 4-1. Sequencing and alignment results from libraries sequenced using target 
enrichment (TE), low coverage whole genome sequencing (Lc-WGS) and whole-
genome sequencing (WGS) ......................................................................................... 164 
 
Table 4-2. Plant hosts and the location of the 101 samples of Pseudoperonospora spp. 
samples used for the population analysesa .................................................................. 166 
 
Table 4-3. Pairwise FST comparisons among Pseudoperonospora cubensis (clade I and 
II) and Pseudoperonospora humuli. ............................................................................. 167 
      

 

vii 

 

 

Table 4-4. Analysis of molecular variance (AMOVA) for Pseudoperonospora spp. 
grouped by host. The significance of variance was tested from 999 permutations of the 
dataa. ............................................................................................................................ 168 
 
Table 4-5. Genotypic diversity estimates for Pseudoperonospora spp. samples grouped 
by host (clade). ............................................................................................................. 169 
 
Table 4-6. Analysis of molecular variance (AMOVA) of Midwest subpopulationsa of 
Pseudoperonospora cubensis collected from Cucumis sativus. The significance of 
variance was tested from 999 permutations of the dataa. ............................................ 170 
 
Table 4-7. Pairwise FST comparisons among subpopulations of Pseudoperonospora 
cubensis collected from C. sativus in the Midwesta. ..................................................... 171 
 
Table 4-8. Genotypic diversity estimates and index of association of 
Pseudoperonospora cubensis subpopulations  collected from C. sativus in the Midwest.
 ...................................................................................................................................... 172 
 
Table 4-9. Analysis of molecular variance (AMOVA) of Michigan subpopulations of 
Pseudoperonospora humuli. The significance of variance was tested from 999 
permutations of the dataz. ............................................................................................ 173 
 
Table 4-10. Pairwise FST comparisons among subpopulations of Pseudoperonospora 
humuli collected from hop yards in Michigana. ............................................................. 174 
 
Table 4-11.Genotypic diversity estimates and index of association of Michigan 
subpopulations of Pseudoperonospora humuli. ........................................................... 175 
 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

viii 

 

 

LIST OF FIGURES 

 
Figure 2-1. Regression and amplification curves of Pseudoperonospora cubensis and P. 
humuli DNA using qPCR. ............................................................................................... 63 
 
Figure 2-2. Standard curves for the quantification of Pseudoperonospora cubensis and 
P. humuli sporangia using qPCR. .................................................................................. 64 
 
Figure 2-3. Monitoring of Pseudoperonospora cubensis and P. humuli sporangia using 
Burkard spore traps in Ingham (A) and Muskegon (B) in 2018. ..................................... 65 
 
Figure 2-4. Monitoring of Pseudoperonospora cubensis and P. humuli sporangia using 
Burkard spore traps in Berrien (A), Ingham (B) and Muskegon (C) in 2019. ................. 66 
 
Figure 3-1. Location of spore traps by county in 2018 (A) and 2019 (B). Green and blue 
dots indicate Burkard and impaction spore traps, respectively. ................................... 110 
 
Figure 3-2. Standard curves for the quantification of Pseudoperonospora cubensis clade 
I, P. cubensis clade II and P. humuli DNA using the qPCR protocol A. ....................... 111 
 
Figure 3-3. Standard curves for the quantification of Pseudoperonospora cubensis clade 
I, P. cubensis clade II and P. humuli DNA using the qPCR protocol B ........................ 112 
 
Figure 3-4. Linear regression of Cq values as a function of sporangial numbers. ....... 113 
 
Figure 3-5. Logistic regression of qPCR results from Burkard (A) and impaction (B) 
spore traps as a function of sporangial numbers. ........................................................ 114 
 
Figure 3-6. Monitoring of Pseudoperonospora cubensis and P. humuli sporangia using 
spore traps in Muskegon (A) and Bay (B) counties in 2018 (Location B). .................... 115 
 
Figure 3-7. Monitoring of Pseudoperonospora cubensis and P. humuli sporangia using 
spore traps in Allegan (A) and Saginaw (B) counties in 2018. ..................................... 117 
 
Figure 3-8. Monitoring of Pseudoperonospora cubensis and P. humuli sporangia using 
spore traps in Muskegon (A) and Bay (B) counties in 2019. ........................................ 119 
 
Figure 3-9. Monitoring of Pseudoperonospora cubensis and P. humuli sporangia using 
spore traps in Allegan (A) and Saginaw (B) counties in 2019. ..................................... 121 
 
Figure 3-10. Monitoring of P cubensis and P. humuli sporangia using spore traps in 
Ingham (A) and Berrien (B) counties in 2019. .............................................................. 123 
 
Figure 3-11. Standard curves for the quantification of Pseudoperonospora cubensis and 
P. humuli sporangia using the qPCR protocol A. ......................................................... 125 

 

ix 

 

 

Figure 3-12. Standard curves for the quantification of Pseudoperonospora cubensis 
clade I, P. cubensis clade II and P. humuli sporangia using the qPCR protocol B. ..... 126 
 
Figure 4-1. Cucumis sativus planted acreage and number Humulus lupulus planted in 
Michigan by county (Adapted from Neufeld, 2017). ..................................................... 176 
 
Figure 4-2. Depth coverage of high-quality SNPs across libraries. .............................. 177 
 
Figure 4-3. Characterization of sequencing error among technical replicates. ............ 178 
 
Figure 4-4. Genetic differentiation of Pseudoperonospora cubensis and 
Pseudoperonospora humuli samples. .......................................................................... 179 
 
Figure 4-5. Ordination plots of Pseudoperonospora spp. based on 2,978 SNPs. ........ 180 
 
Figure 4-6. Frequency and geographic distribution of Pseudoperonospora cubensis and 
Pseudoperonospora humuli genotypes (MLG). ............................................................ 181 
 
Figure 4-7. The relationship between genetic differences among samples and 
geographic distances of the locations from which samples originated. ........................ 182 
 
Figure 4-8. Estimation of the degree of linkage disequilibrium within 
Pseudoperonospora cubensis and Pseudoperonospora humuli of Michigan. ............. 183 
 

 

 

 

 

x 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CHAPTER I: LITERATURE REVIEW - PSEUDOPERONOSPORA CUBENSIS THE 

CAUSAL AGENT OF CUCURBIT DOWNY MILDEW 

1 

 

 

INTRODUCTION 
 

Pseudoperonospora cubensis (Berk. & M. A. Curtis) Rostovzev, the causal agent 

of cucurbit downy mildew (CDM), is a highly destructive pathogen recognized as one of 

the greatest contemporary disease threats to cucumber production worldwide 

(Brzozowski et al., 2016). In the U.S., two distinct evolutionary clades of P. cubensis 

(clade I and II) affect the production of cucurbit species including agronomically 

important crops such as cucumber, pumpkin, watermelon, and squash (Wallace et al., 

2020). The disease only affects foliage creating small, irregular to angular chlorotic 

areas, and in most cases, sporulation on the lower leaf surface (Cohen et al., 2015). 

Leaf lesions coalesce and become necrotic leading to leaf blighting and premature 

defoliation which results in stunted plants and yield reduction, especially in cucumber 

(Cucumis sativus L.) and squash (Cucurbita moschata) (Reuveni et al., 1980; Adams et 

al., 2019; Hausbeck et al., 2019; Perla et al., 2019).  

In the U.S., the disease was only of minor concern prior to 2004 due to the 

deployment of resistant cucumber cultivars starting in 1960 (Thomas et al., 2017c). 

However, in 2004, a highly virulent strain of P. cubensis was introduced into U.S. that 

overcame this host resistance and stunned the cucumber industry (Holmes et al., 2014; 

Thomas et al., 2017a). In the absence of fully resistant cucurbit cultivars, chemical 

control is currently the most effective strategy to control CDM (Hausbeck and 

Goldenhar, 2017) but the evolution of resistant P. cubensis isolates to multiple 

fungicides has created an urgent need for alternative disease management strategies 

(Thomas et al., 2017).  

 

2 

 

 

Public breeding programs have made progress in the development of new 

cultivars to respond to the rising virulence of CDM, and have released slicing 

cucumbers with good levels of resistance (Brzozowski et al., 2016). However, the 

resistance of these varieties to the downy mildew pathogen has not yet been 

incorporated into processing (pickling) cucumbers that represent a significant portion of 

the cucurbit crops grown in the country. The U.S. is the seventh-largest producer of 

cucumber in the world (Keinath et al., 2017) and in 2019, 756,000 metric tons of fresh 

market (20%) and pickling cucumbers (80%) were grown on 42000 ha for a combined 

value of $279 million (USDA, 2020). Michigan is the second-largest producer of 

cucumbers in the country where approximately half of the national production of 

cucumbers (300,000 metric tons fresh and pickling) is harvested every year (USDA, 

2020). Similarly, Michigan is also a big producer of other susceptible crops to CDM such 

as squash and pumpkin. In 2019, 85,600 and 41,200 metric tons of squash and 

pumpkins, respectively, were grown on a combined area of 5700 ha (USDA, 2020). 

P. cubensis clade II is especially destructive on cucumbers (Holmes et al., 2014). 

The pathogen cannot overwinter in northern regions of the U.S. that experience frost, 

but the disease reoccurs yearly due to the influx of airborne sporangia from 

overwintering sites with mild winters (Ojiambo and Holmes, 2010; Quesada-Ocampo et 

al., 2012). CDM has changed the way cucumber growers manage their fields in 

Northern states such as Michigan. The pickling cucumber growers of Michigan apply 

fungicides frequently to limit CDM, with an estimated cost of more than $6 million 

annually (Goldenhar and Hausbeck, 2019). Oxathiapiprolin (FRAC 49), ethaboxam 

(FRAC 22), fluazinam (FRAC group 29), cyazofamid (FRAC 21), 

 

3 

 

 

ametoctradin/dimethomorph (FRAC 45/40), and the broad-spectrum fungicides 

mancozeb (FRAC M03) and chlorothalonil (FRAC M05) are among the best fungicide 

alternatives for CDM control in different regions of the U.S. (Goldenhar and Hausbeck, 

2019; Adams et al., 2020; Dutta, 2020; Miller et al., 2020).  

Application timing of fungicides is crucial when weather conditions are especially 

favorable for CDM and to reduce the risk of the pathogen developing resistance, 

fungicides must be applied preventively (Hollomon, 2007). To optimize application 

timing, the airborne concentration of P. cubensis sporangia can be monitored to 

coordinate the initiation of fungicide applications with the arrival of the pathogen into 

cucumber production fields (Granke et al., 2013; Goldenhar and Hausbeck, 2019). The 

accurate detection and quantification of sporangia can enable a more efficient 

application timing of fungicides and possibly, delay the development of pathogen 

resistant isolates to fungicides.  

Spore trapping can provide quantitative data on airborne spore numbers (Dung 

et al., 2018); however, the processing and microscopic examination of spore trap 

samples is time consuming and can result in misidentification due to morphological 

similarities between Pseudoperonospora taxa (e.g. P. cubensis and P. humuli). The 

combination of spore traps and PCR technologies is an ongoing area of investigation 

that seeks to improve and accelerate the detection of P. cubensis sporangia from air 

samples. Next generation sequencing technologies have recently facilitated the 

identification of new diagnostics markers for pathogen detection (Withers et al., 2016; 

Rahman et al., 2019) and the development of qPCR assays that could improve the 

monitoring of airborne P. cubensis sporangia (Summers et al., 2015a).   

 

4 

 

 

Similarly, these technologies have also facilitated the implementation of new 

tools to monitor CDM populations (Summers et al., 2015; Wallace & Quesada-Ocampo, 

2017; Withers et al., 2016). This important progress has made possible the execution of 

more comprehensive genetic studies to better understand the epidemiology of CDM in 

the U.S. (Summers et al., 2015; Thomas et al., 2017). This review briefly summarizes 

our current understanding of P. cubensis biology including taxonomy, dispersal, 

management, fungicide resistance and population genetics. Additionally, we also 

elaborate on future directions for the effective control of CDM including plant breeding 

and early pathogen detection 

TAXONOMY AND DISEASE CYCLE 
 

Pseudoperonospora cubensis (Berk. & Curt.) Rost., the causal agent of cucurbit 

downy mildew (CDM), is an oomycete plant pathogen belonging to the family 

Peronosporaceae (Oomycota, Oomycetes, Peronosporales). This family is made up of 

an extensive number of plant pathogens that threaten natural and managed ecosystems 

including all downy mildews (DM) and other genera such as Phytopythium spp., 

Halophytophthora spp. and  Phytophthora spp. (Thines and Choi, 2015). All DM are 

considered obligate biotrophic plant parasites and as such, they can only grow in 

association with living host tissue (Thines, 2014). Phytophthora spp., on the other hand, 

are hemibiotrophic or necrotrophic and only a small group of species are obligate 

biotrophs (Thines, 2014; Bourret et al., 2018).  

So far, 19 downy mildew genera have been described that contain more than 

700 species (Thines and Choi, 2015). Most DM genera are present in three major 

monophyletic groups: 1) downy mildews with colored conidia (Peronospora and 

 

5 

 

 

Pseudoperonospora), 2) downy mildews with pyriform haustoria (Basidiophora, Benua, 

Bremia, Novotelnova, Paraperonospora, Plasmopara, Plasmoverna, and Protobremia), 

and the brassicolous downy mildews (Hyaloperonospora and Perofascia). Apart from 

these three groups, several grass-parasitic downy mildew genera have been described, 

of which the graminicolous downy mildews with lasting sporangiophores 

(Graminivora, Poakatesthia, and Viennotia) seem to be monophyletic (Thines, 2014; 

Thines and Choi, 2015; Bourret et al., 2018). DM were thought to form a single 

monophyletic group, however, a recent multigene phylogenetic analysis showed that 

Graminicolous downy mildews (GDM), brassicolous downy mildews (BDM) and downy 

mildews with colored conidia (DMCC) form a monophyletic clade with the 

Phytophthora taxon totara; while downy mildews with pyriform haustoria (DMPH) were 

placed in their own clade (Bourret et al., 2018). 

The genus Pseudoperonospora has been place alongside Peronospora in the 

group of downy mildews with colored conidia. Six species have been described within 

the Peronosporaceae genus, including two economically important species: P. cubensis 

and P. humuli (causal agent of hop downy mildew). P. cubensis causes disease on 

approximately 20 cucurbits genera (approximately 40 – 60 different species) including 

cucumber (Cucumis sativus), cantaloupe (C. melo), pumpkin (Cucurbita maxima), 

watermelon (Citrullus lanatus), squash (Cucurbita pepo), gourd (C. moschata), and wild 

cucurbit species such as balsam apple (Momordica balsamina ), bitter melon (M. 

charantia) and Buffalo Gourd (Cucurbita foetidissima) (Quesada-Ocampo et al., 2012; 

Savory et al., 2011; Wallace et al., 2014; Wallace et al., 2015).  

 

6 

 

 

P. cubensis requires a living host to complete its life cycle which begins when 

sporangia (2n) land on the adaxial surface of susceptible host. Under the right 

environmental conditions (15-20 °C, 1-5h of leaf wetness), sporangia release zoospores 

(2n) capable of swimming and encysting in host stomata (Granke et al., 2013; Cohen et 

al., 2015). From the encysted zoospore, a germ tube forms, penetrating the host tissue 

through an appressorium. Hyphae (2n) grow into the intracellular space where nutrient 

acquisition occurs. Sporangiophores form clonally from the intracellular growing hyphae 

holding sporangia produced mitotically in their tips (2n). Sporangia dislodge from 

Sporangiophores by a twisting mechanism that occurs when relative humidity (RH) 

decreases, and then fly through the air until they reach a new susceptible host re-

initiating the cycle. 

Several oomycete pathogens reproduce in a mixed mode (asexual and sexual 

reproduction), however, it is not clear if P. cubensis undergoes a sexual phase in U.S. 

fields. The formation of oospores by P. cubensis seems to be rare but it has been 

reported several times in other countries such as China, Israel and India (Cohen et al., 

2003; Savory et al., 2011; Zhang et al., 2012). A recent study confirmed the presence of 

two different mating types in the U.S. (A1 and A2) as well as its ability to produce viable 

oospores in vitro. (Thomas, A. Carbone, I. Ojiambo, 2013).  This suggests that P. 

cubensis could potentially reproduce sexually in cucurbits within U.S. but the actual 

frequency of oospore formation could be very low due to the strict association of each 

mating type to a particular host (i.e. the A1 mating type isolates were uniquely found in 

cucumbers while the A2 mating type was mainly found in squash) (Thomas et al., 2016; 

Cohen et al, 2011; Cohen et al., 2015). The incidence of sexual reproduction and the 

 

7 

 

 

formation of oospores could have an important role in the epidemiology of CDM in the 

country. Monitoring of sexual reproduction in U.S. fields is key to identify new sources of 

genetic variation.  

SURVIVAL AND DISPERSAL OF P. CUBENSIS 
 

P. cubensis sporangia do not survive for a long time on non-living or necrotic 

tissue and its ability to infect cucumbers is reduced greatly under dry conditions, 

surviving only for 22 hours at temperature between 35 and 40 Celsius degrees (Cohen 

and Rotem, 1971). Additionally, overwintering-sexual structures such as oospores have 

not been detected in soil or host tissue in the U.S., and are not considered a source of 

inoculum (Naegele et al., 2016). Thus, P. cubensis survival in the U.S. depends on the 

availability of susceptible hosts (Savory et al., 2011) and the pathogen cannot survive in 

fields of regions where cucurbits cannot be grown year-round due to long winters with 

frost. In northern U.S., CDM reoccurs yearly due to the influx of sporangia that 

originated from warm weather plantings along the eastern seaboard and/or 

greenhouses in colder locales where the pathogen can survive on living host tissue 

(Ojiambo and Holmes, 2010; Quesada-Ocampo et al., 2012). CDM outbreaks in the 

great lakes region of the U.S. and Ontario, Canada are thought to result from the arrival 

of sporangia from southern states (e.g. Florida) that migrate north by wind currents 

using plantings of susceptible crops (Ojiambo et al., 2015). However, migration 

inferences made at a genetic level suggest a more restricted movement of sporangia in 

the country (Quesada-Ocampo et al., 2012). 

Although P. cubensis sporangia can travel up to 1,000 kilometers via air currents 

and migrate between fields in different states (Ojiambo and Holmes, 2010; Ojiambo et 

 

8 

 

 

al., 2015; Naegele et al., 2016), the CDM population from southern U.S. (i.e. Florida, 

Georgia, North Carolina) are highly dissimilar to the pathogen population in the Upper 

Midwest (Quesada-Ocampo et al., 2012). Evidence has been found that support the 

exchange of migrants among states in the upper Midwest and Canada (Naegele et al., 

2016) but evidence of the movement of sporangia between northern and southern 

states has not yet been found.   

The progressive movement of the pathogen between states depends mainly on 

three factors: the asexual production of sporangia, its passive atmospheric 

transportation and the availability of new susceptible host (Ojiambo & Holmes, 2010). 

These factors in combination with environmental variables are correlated with the 

occurrence of CDM (Granke et al., 2013) and provide the basis of the CDM forecasting 

system (CDM ipmPIPE) that estimates the risk of outbreaks at any particular area 

(Ojiambo et al., 2015). Under field conditions, airborne sporangial concentrations, time 

post-planting, temperature, and leaf wetness are positively associated with disease 

occurrence, while solar radiation is the only factor negatively associated with disease 

(Granke and Hausbeck, 2011; Granke et al., 2013). 

Recent studies indicate that the P. cubensis can also be transmitted by seeds 

(Cohen et al., 2014) and infect wild cucurbits (Wallace et al., 2014; Wallace et al., 2015; 

Wallace and Quesada-Ocampo, 2016). CDM symptoms and sporulation have been 

observed on the leaves of wild species such as Balsam apple (Momordica balsamina), 

bitter melon (Momordica charantia), buffalo gourd (Cucurbita foetidissima), and bottle 

gourd (Lagenaria siceraria) (Wallace and Quesada-Ocampo, 2016), however, it is still 

unknown if the pathogen can overwinter in the dormant tissue of these species (Wallace 

 

9 

 

 

et al., 2014; Wallace et al., 2015). Further research is needed to establish the role of 

non-commercial cucurbits in the yearly CDM epidemic, which will aid the efforts of the 

CDM ipmPIPE to predict disease outbreaks. 

GENETIC STRUCTURE 
 

Research on P. cubensis populations from Europe and the U.S. have identified 6 

distinct genetic clusters among 465 samples (Quesada-Ocampo et al., 2012; Kitner et 

al., 2015). Some clusters were more frequently associated with particular geographical 

regions, however, all of them were detected in Europe and in the U.S. This suggests 

that some genotypes are persistent and widely dispersed and/or have migrated from 

one population to others (Quesada-Ocampo et al., 2012). Initial studies may have 

underestimated the diversity of the populations due to the low number of markers used 

(Quesada-Ocampo et al., 2012; Kitner et al., 2015), however, the general structure 

patterns have also been observed in analyses using  larger numbers of genetic markers 

(Summers et al., 2015b; Thomas et al., 2017a; Wallace and Quesada-Ocampo, 2017).      

At the genetic and phenotypic level (i.e. host preference), P. cubensis is 

structured by host in the U.S. (Thomas et al., 2017a; Wallace et al., 2020). In fact, it was 

recently shown that  P. cubensis in the U.S. can be divided  into two host-specific 

clades (Wallace et al., 2020). Further genetic studies have confirmed that these two 

clades are host-adapted at the cucurbit species level (Summers et al., 2015b; Thomas 

et al., 2017a; Wallace et al., 2020), with clade I isolates recovered more frequently from 

commercial varieties of Cucurbita pepo, C. moschata, C. maxima, and Citrullus lanatus 

and clade II isolates associated more frequently with commercial varieties of the 

Cucumis sativus and Cucumis melo (Wallace et al., 2020). Additionally, clade II isolates 

 

10 

 

 

were also found infecting the wild cucurbit species Lagenaria siceraria while clade I was 

also isolated from the wild species Momordica balsamina and Momordica charantia 

(Thomas, A. Carbone, I. Ojiambo, 2013; Wallace et al., 2020). Both clades were only 

found with low frequency in Cucumis melo and Cucurbita foetidissima (Wallace et al., 

2020).  

It is still unclear if sexual reproduction occurs under field conditions within the P. 

cubensis population of the U.S. However, signs of recombination were found using 

genetic markers in clade I consistent with a sexually reproducing population, while no 

evidence of random mating was found for clade II (Wallace et al., 2020). This suggests 

that only clade I could be heterothallic while clade II may only reproduce clonally. 

Thomas et al., (2017) confirmed the presence of two mating types (A1 and A2) in the 

U.S. able to form oospores in vitro but information on the clade membership of the 

isolates used is not available. The existence of two different mating types within each 

clade has not been confirmed and it is unknown if recombination between isolates of 

different clades can occur. However, both clades are rarely detected within a single host 

(Cucumis melo) suggesting the possibility of two different mating-types from each clade 

encountering each other is low.  

As well as clustering by host, clustering between P. cubensis isolates by 

geographic location has also been reported in the U.S. (Quesada-Ocampo et al., 2012; 

Naegele et al., 2016) but it is unclear if this geographic structure is real or is an artifact 

of the population differentiation driven by the host or a temporal effect of the sampling. 

Population studies of P. cubensis have are limited by the sporadic occurrence of the 

disease due to the obligate nature of the pathogen. This makes the collection of isolates 

 

11 

 

 

highly dependent on the availability of susceptible hosts, whose production is regionally 

and temporally affected. The cucurbit growing season along the states in the eastern 

seaboard of the U.S has minimal overlap making comparisons among P. cubensis 

populations from northern and southern states difficult without considering a temporal 

factor. In Northern U.S., cucurbits are mainly grown during the summer while southern 

states such as Florida and Georgia produce cucurbits only during the spring and fall 

(Aerts and Mossler, 2003). Thus, the genetic differentiation in space (e.g. region, state) 

detected previously (Quesada-Ocampo et al., 2012) might be an artifact of the host 

driven structure and/or a biased sampling. A population study performed in Czech 

Republic over two years of sampling revealed no clustering based on geographical 

origin (Kitner et al., 2015). 

PATHOGENIC VARIATION 
 

The change in the host range of CDM that occurred in Europe in 2009, when P. 

cubensis became a significant problem for species like Cucurbita moschata, C. pepo, C. 

maxima and Citrullus lanatus, was attributed to a significant change in the structure of 

pathogen population (Kitner et al., 2015). Pre-epidemic samples were different 

significantly from samples collected after 2009 and they clustered in completely different 

clades. Only a limited number of heterozygous samples collected after 2009 clustered in 

the pre-epidemic clade, suggesting the occurrence of rare recombination events 

between populations (Cohen et al., 2015).  

In the same way, the emergence of new pathotypes (physiological races) has 

been proposed as an explanation for the breaking of host resistance in the United 

States. However, this hypothesis has not yet been proved due to the small number of 

 

12 

 

 

samples collected before 2004, and the limited resolution of the studies performed so 

far (Quesada-Ocampo et al., 2012). A better understanding of P. cubensis populations 

at the local level is key to control and prevent the emergence of more virulent pathogen 

populations with additional levels of fungicide resistance.  

FUNGICIDE RESISTANCE 
 

In the absence of fully resistant cucumber varieties, fungicide use is currently the 

most effective method to control CDM (Hausbeck and Goldenhar, 2017). However, P. 

cubensis is a high-risk pathogen in terms of evolving fungicide resistance both because 

of its shorth generation time on the genetically uniform monocultures of its hosts and the 

huge population size during outbreaks that offer many opportunities for mutations 

(Kitner et al., 2015). In fact, P. cubensis, was the first pathogen to be reported as 

resistant to the phenylamide fungicide mefenoxam (FRAC 4), a widely used fungicide 

against most of the oomycete plant pathogens (Reuveni et al., 1980). Over 17 different 

fungicides (representing 15 FRAC codes) are registered to control CDM (Table 1-1), 

however, complete resistance or reduction in the sensitivity to multiple fungicides has 

been reported within P. cubensis populations (Urban and Lebeda, 2006; Goldenhar and 

Hausbeck, 2019).  

Complete resistance of P. cubensis to fungicides in the FRAC groups 4 

(phenylamides), 11 (quinone outside inhibitors), and 40 (carboxylic acid amides) (Gisi 

and Sierotzki, 2015; Ojiambo et al., 2015) has been documented. In the U.S., single-site 

fungicides including mefenoxam (FRAC 4) and azoxystrobin (FRAC 11) were ineffective 

when the pathogen reemerged in 2004 (Ernest et al., 2005; Gevens and Hausbeck, 

2005; Thornton et al., 2006). Since that time, P. cubensis resistance to dimethomorph 

 

13 

 

 

(FRAC 40) (Zhu et al., 2007) and mandipropamid (FRAC 40) (Hausbeck and Cortright, 

2010; Blum et al., 2011) has been reported in the U.S. (Holmes et al., 2014; Keinath, 

2015). Similarly, reduced efficacy of fluopicolide and cymoxanil (FRAC 43) against CDM 

has been observed in field trials in Michigan (Hausbeck and Linderman, 2014; 

Goldenhar and Hausbeck, 2019), Georgia (Langston and Sanders, 2013), and North 

Carolina (Adams and Quesada-Ocampo, 2014; Keinath, 2015). Propamomcarb (FRAC 

28) was effective against CDM for several years, but in 2013 its efficacy was 

compromised in field trials in North Carolina (Keinath, 2015; Thomas et al., 2018), 

Pennsylvania (Gugino and Grove, 2016), and Michigan (Hausbeck and Linderman, 

2014; Hausbeck et al., 2017).  

The number of effective fungicides against CDM is limited, and fungicide efficacy 

can vary between years (Goldenhar and Hausbeck, 2019) due to changes in 

environmental conditions and the pathogen population. However, oxathiapiprolin (FRAC 

49), ethaboxam (FRAC 22), fluazinam (FRAC group 29), cyazofamid (FRAC 21), 

ametoctradin/dimethomorph (FRAC 45/40), and the broad-spectrum fungicides 

mancozeb (FRAC M03) and chlorothalonil (FRAC M05) have shown a good control 

levels of CDM in field trials in Ohio (Miller et al., 2020), Georgia (Dutta, 2020), North 

Carolina (Adams et al., 2020) and Michigan (Goldenhar and Hausbeck, 2019; Hausbeck 

et al., 2019) 

Oxathiapiprolin is a relatively new active ingredient with proven efficacy against 

P. cubensis (Cohen 2015; Goldenhar and Hausbeck 2016). However, it is a single site 

fungicide inhibitor and as such, it is considered at high risk of pathogen resistance 

(Cohen 2015; FRAC 2018). Similarly, ethaboxam had demonstrated efficacy against 

 

14 

 

 

CDM (Quesada-Ocampo and Adams, 2014; Gugino and Grove, T.L., 2020) and was 

classified as a fungicide for which the development of resistance is at low to medium 

risk (FRAC, 2020); still, in 2017 its efficacy was compromised in field trials in Michigan 

(Goldenhar and Hausbeck, 2019). Fluazinam, cyazofamid and 

ametoctradin/dimethomorph have shown consistent good control of CDM (Adams and 

Quesada-Ocampo, 2014; Keinath, 2015; Goldenhar and Hausbeck, 2019) and only 

Fluazinam is not widely used  by cucurbit growers due to its high cost (Neufeld et al., 

2017). Resistance to broad spectrum fungicide is rare in P. cubensis populations, but 

metalaxyl-resistant isolates found in Israel also exhibited moderate levels of resistance 

against active ingredients such as mancozeb (Reuveni et al., 1980) 

 It seems inevitable to avoid hastening the emergence of resistance given the 

limited amount of options against CDM and the high rates at which fungicides are 

applied (every 7 to 10 days). Thus, resistance management is key to maintaining the 

efficacy of single-site fungicides, and growers are encouraged to rotate among 

fungicides from different FRAC groups (FRAC, 2020). The continuous monitoring of 

fungicide efficiency at the local level is essential to provide the best recommendations to 

cucumber growers in the U.S. and ensure that P. cubensis is effectively managed.

 

15 

 

 

Table 1-1. List of fungicides registered for the control of cucumber downy mildew in the U.S.  

Product name  Registrant 

FRAC 
code 

a Reduce 
efficacy 

FRAC chemical 

group 

Target site 

Yes 

No 

PhenylAmides 
(PA) - fungicides  RNA polymerase I 
Multi-site contact 

Chloronitrites 

activity 

Yes 

Benzamides 

Yes 

Carbamates 

Delocalization of 

spectrin-like 

proteins 

Cell membrane 
permeability, fatty 
acids (proposed) 

No 

No 

No 

Yes 

Quinone inside 
inhibitor (Qil) - 

fungicides 

Dithio-

carbamates 

(M03) / thiazole 
carboxamide (22) 
Cyanoacetamide-

oxime 

No 

Unestablished 

Complex III: 

cytochrome bc1 at 

Qi site 

Multi-site contact 
activity (M03) / B-
tubulin assembly in 

mitosis (22) 

Unknown 

Uncouplers of 

oxidative 

phosphorylation 

Active 
ingredient 

Metalaxyl/ 
mefenoxam 

Ridomil 

Syngenta 

   4 

Chlorothalonil 

Bravo Weather 

Stik 

Syngenta 

M05 

Fluopicolide 

Presidio 

Valent 

Propamomcarb  Previcur Flex 

Bayer 

Cyazofamid 

Ranman 

FMC 

43 

28 

21 

Zoxamide/ 
mancozeb 
Mancozeb 

Cymoxanil 

Gavel 

Dithane 

Curzate 

Gowan 

22/M03 

BASF 

M03 

DuPont 

27 

29 

Fluazinam 

Omega 

Syngenta 

    aReduce efficacy or completed resistance reported in the U.S. 
 

 

 

16 

 

    Table 1-1. (cont’d)  

Active 
ingredient 

Mandipropamid 

Dimethomorph 
Ametoctradin/ 
dimethomorph 
Fluxapyroxad/ 
pyraclostrobin 
Pyraclostrobin 
Famoxadone/ 
cymoxanil 

Product 
name 

Revus 

Forum 

Zampro 

Priaxor 

Cabrio 

Tanos 

Registrant 

Syngenta 

BASF 

BASF 

BASF 

BASF 

40 

40 

45/40 

11/7. 

11 

DuPont 

11/27. 

Oxathiapiprolin 

Orondis 

Syngenta 

49 

Ethaboxam 

Elumin 

Valent 

22 

      aReduce efficacy or completed resistance reported in the U.S

 

 

17 

FRAC 
code 

a Reduce 
efficacy 

FRAC chemical 

group 

Target site 

Yes 

Yes 

No 

Yes 

Yes 

Yes 

No 

No 

Carboxylic Acid 
Amides (CAA) - 
fungicides (40) 

Cellulose synthase 
(40)/ complex III: 

cytochrome bc1 (45) 

Quinone outside 
Inhibitors (QoI) - 
fungicides (11) 

OSBPI oxysterol 
binding protein 

homologue 
inhibition 

Complex III: 

cytochrome bc1 at 

Qo site (11) 

Lipid homeostasis 
and transfer/storage 

Thiazole 

carboxamide 

ß-tubulin assembly in 

mitosis 

 

 
  
 
 

 

 

 

 

BREEDING EFFORTS 
 

Cucumber (Cucumis aestivus) is a widely cultivated plant of the Cucurbitaceae 

family, with annual production above 71 million tons globally (FAO, 2013). The U.S. is 

the fifth largest producer of cucumber, and in 2019, 756,000 metric tons of fresh market 

and pickling cucumbers were grown on approximately 42000 ha for a combined value of 

$279 million (USDA, 2020). Several diseases affect cucumber production including 

target spot,  powdery mildew and Phytophthora crown and root rot (Tuttle McGrath, 

2004; Keinath et al., 2017), however, cucurbit downy mildew (CDM) caused by the 

oomycete, Pseudoperonospora cubensis is probably the most important disease of 

cucumber in the country (Thomas et al., 2017a). CDM is a highly destructive foliar 

disease able to cause 50% reduction in cucumber yield even after fungicides are 

applied one-week post-symptom appearance (Cohen et al., 2015). 

Currently no cultivar has robust resistance to the disease, but for decades, 

downy mildew on cucumbers was effectively managed with genetic host resistance. 

(Sitterly, 1972). The introgression of the dm-1 gene from PI 197087 into the cultivars 

‘Polaris’, ‘Poinsett’, ‘Pixie’, and ‘Chipper’ provided cucumber growers with genetic 

control of downy mildew for more than 40 years (Call et al., 2012a; Cohen et al., 2015; 

Thomas et al., 2017a). From 1961 to 2003, downy mildew was only a moderate problem 

in North America and was easily controlled with fungicides (Cohen et al., 2015). 

However, the resistance of commercial cultivars in the U.S. was defeated in 2004 when 

a highly virulent strain of P. cubensis was introduced in the country (Thomas et al., 

2017a). Since then, P. cubensis has reoccurred yearly in cucumber production areas of 

 

18 

 

 

eastern U.S., costing cucumber growers millions annually in fungicide sprays to protect 

their crop (Granke and Hausbeck, 2011; Savory et al., 2011).  

Over the last ten years, a number of new downy mildew-resistant cucumber lines 

have been released by public and private breeding programs including the slicing 

varieties SV3462CS, SV4719CS, and Bristol, and the pickling varieties Citadel and 

Peacemaker from Seminis (Holdsworth et al., 2014; Brzozowski et al., 2016). These 

varieties have shown an intermediate level of resistance to P. cubensis and reached 

comparable yields to commercial standards (e.g. Vlaspik) but still required the utilization 

of fungicides to provide full disease control (Hausbeck and Goldenhar, 2017). The 

Cornell breeding program have made significant progress in the development of new 

cultivars to control CDM releasing slicing cucumbers with exceptional levels of 

resistance (Brzozowski et al., 2016). The released line “DMR-NY401” retained the 

disease resistance of its predecessor line “DMR-NY264” while showing significantly 

higher fruit length, yield, and earliness of initial harvest (Brzozowski et al., 2016). 

However, the resistance of these varieties has not yet been incorporated in pickling 

cucumber varieties which are highly important in the country.  

Race-specific R gene breeding: Plant pathogens such as P. cubensis secrete 

effector proteins during infection that modulate host innate immunity (Goss et al., 2013). 

Many of these effectors function only as virulence factors, but others can be recognized 

by plant R proteins resulting in the activation of effector triggered immunity. In such 

cases, the effectors are known as avirulence (AVR) factors (Brzozowski et al., 2016). In 

most cases, the response induced by AVR factors involves the hypersensitive response 

(HR) followed by restriction of the invading pathogen (Vleeshouwers et al., 2008).  

 

19 

 

 

Virulence factor can be identified throughout the phenotypic characterization of 

the P. cubensis population and can then be functionally profiled on cucurbit species to 

detect cognate R genes. R genes have been successfully identified using virulence 

factors in the model system P. infestans-potato plant and other fungal species such as 

Cryptococcus neoformans (Vleeshouwers et al., 2008; Desjardins et al., 2017). 

However, the identification of R genes in the cucurbit downy-mildew system using this 

strategy faces two important obstacles. First, as an obligate biotroph, P. cubensis 

requires a living host tissue for reproduction and dispersal, complicating the 

maintenance, identification and phenotyping of isolates. Secondly, the specific spectrum 

of novel CDM R-genes cannot be assessed in the absence of a diverse panel of 

isolates or pathotypes (isolates with the same pathogenicity). Due to these difficulties 

almost every phenotypic or genotypic study performed so far includes only a limited 

number of isolates.  

An effort to identify diagnostic pathotypes of P. cubensis was recently performed 

by Thomas, et al. (2017). In this study, thirteen different pathotypes were identified 

based on a set of 15 different cucurbit species. The authors suggest the existence of 10 

different avirulence factors (Avr genes), assuming that disease resistance is only 

expressed when an R gene product in the host can recognize the pathogen’s 

corresponding effector. Interestingly, specific pathotypes form subgroups according to 

mating type which suggests the association between virulence and mating type 

(Thomas et al., 2017c).   

The ways in which virulence factors evolve will be highly valuable in breeding of 

specific R genes and the designing of effective management strategies in agricultural 

 

20 

 

 

systems.  Unfortunately, these concepts are poorly understood for obligate plant 

pathogen (Thomas et al., 2017c). In order to understand the evolution of Avr genes, 

particularly in the cucurbit downy-mildew system, substantial efforts need to be made to 

identify virulence factors and their genes. This could help to survey the temporal and 

spatial distribution of virulence factors and race structure of the P. cubensis population.  

EARLY DETECTION OF P. CUBENSIS 
 

In the absence of fully resistant cucurbit cultivars, chemical control is currently 

the most effective method for controlling CDM (Wu et al., 2016). However, the risk for 

the evolution of resistance by P. cubensis is high due to its fast mode of reproduction, 

and the low abundance of multi-site inhibitors to control the disease. It seems inevitable 

to avoid hastening the emergence of resistant P. cubensis isolates given the limited 

amount of options for CDM control. Therefore, the reduction of fungicide use and the 

alternation between fungicides in different FRAC groups is key to maintaining the 

efficacy of single-site inhibitors (Goldenhar and Hausbeck, 2019).   

Application timing is crucial to optimize fungicide utilization and preventive 

application of fungicides can reduce the risk of the pathogen developing resistance 

(Hollomon, 2007). Thus, coordinating the initiation of fungicide applications with the 

arrival of pathogens into production fields could result in more efficient fungicide use. In 

California, measurements of aerial spore loads obtained from spore traps have been 

used to schedule the timing of fungicide applications against the lettuce pathogen, 

Bremia lactucae. This strategy resulted in the reduction of sprays without a significant 

increase in disease incidence (Dhar et al., 2019).  

 

 

21 

 

 

Monitoring the airborne concentration of P. cubensis sporangia using spore traps 

could also help to improve the scheduling of fungicides applications to control CDM in 

regions where disease occurrence depends on the influx of P. cubensis sporangia 

(Granke and Hausbeck, 2011; Granke et al., 2013; Dung et al., 2018). However, the 

processing and examination of spore trap samples need improvement to reduce the 

processing time and avoid misidentification. The screening of spore trap samples using 

qPCR can significantly improve the detection of airborne plant pathogens. qPCR has 

significantly improve the sensitivity and specificity for the detection of  Peronospora 

effusa (Klosterman et al., 2014), Bremia lactucae (Kunjeti et al., 2016), and Botrytis 

cinerea (Kunjeti et al., 2016) in spore trap samples. 

The use of molecular markers to differentiate between P. cubensis and P. humuli 

sporangia could aid to improve the airborne monitoring of CDM.  Previous PCR assays 

have been used to monitor airborne Pseudoperonospora humuli sporangia near hop 

yards to inform the timing of fungicide sprays, but since this assay also detects P. 

cubensis, its accuracy decreases in areas where hops and cucurbits are grown in close 

proximity (Gent et al., 2009). The specificity for the detection of P. cubensis was 

improved with the development of a qPCR assay able to differentiate between P. 

cubensis and P. humuli based on the recognition of a single SNP in the cox2 gene 

(Summers et al., 2015a). However, this assay is unable to differentiated between the 

two host adapted clades of P. cubensis and its specificity was compromise in samples 

containing samples from both species (P. cubensis and P. humuli). A new multiplex 

qPCR assay has been designed based on the recognition of a more polymorphic 

mitochondrial loci that also allow the differentiation between P. cubensis and P. humuli 

 

22 

 

 

(Crandall, 2020). These loci have less similarity between species and also allow the 

differentiation between two distinct clades of P. cubensis (clade I and II) (Thomas et al., 

2017a; Wallace et al., 2020).  

In northern regions of the U.S., CDM occurs annually due to an influx of airborne 

sporangia from overwinter sources. Burkard spore traps couple with light microscopy 

have been used to monitor the influx of Pseudoperonospora spp. sporangia and inform 

growers that the disease is likely to occur. Burkard spore traps collect spores by 

vacuuming air into a collection chamber that contains a reel mounted in a clockwork 

mechanism were spores and other particles are impacted onto a greased tape that 

covers the reel (Burkard Manufacturing Co. Ltd., U.K.). Burkard spore traps can 

continuously collect data for up to 7 days and the collecting tape can be split at different 

intervals allowing the precise quantification of spores per unit of time. Although Burkard 

spore traps are the most common devices used to monitor the air concentration of plant 

pathogens, the impaction spore traps have become popular in the last decade (Jackson 

and Bayliss, 2011). These devices possess rods coated with adhesive material, which 

spin at a standard rate to impact and collect the airborne inoculum (TSE Systems, 

Chesterfield, MO), however, impaction traps require constant monitoring to allow 

accurate estimations per unit of time (i.e. an hour or day). Burkard spore traps are 

robust and highly autonomous but impaction spore traps can be more cost-effective and 

easier to use by growers (Jackson and Bayliss, 2011; Choudhury et al., 2016a).  

There are numerous comparative studies investigating the efficiency of these two 

spore traps, and in most cases, Burkard spore traps perform better than impaction 

spore traps. However, impaction spore traps have been used successfully for airborne 

 

23 

 

 

detection of pathogens such as Bremia lactucae, Peronospora effusa and Peronospora 

schachtii (Jenkyn, 1974; Li and Lin, 1999; Evenhuis et al., 2003), and could also be 

used for early detection P. cubensis sporangia in cucumber growing regions.  

 

 

 

 

 

 

 

 

 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

 

24 

 

 

 
 

 
 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

 

 

 

 

 

 

LITERATURE CITED 

25 

 

 

LITERATURE CITED 

 
 
Aerts, M., Mossler, M. (2003).  Crop profile for cucumbers in Florida. In the National IPM 
Database [Data file]. Available from 
https://ipmdata.ipmcenters.org/documents/cropprofiles/FLcucumber2003.pdf 
 
Adams, M. L., and Quesada-Ocampo, L. M. (2014) Evaluation of fungicides for control 
of downy mildew on cucumber. Kinston 2013. Plant Dis Manag Rep 8: V240 
 
Adams, M., Collins, H., Quesada-Ocampo, L.M. (2020) Evaluation of fungicides and 
cultivars for control of downy mildew on cucumber, Kinston, NC 2019. Plant Dis Manag 
Rep Report 14: V107 
 
Adams, M., D’Arcangelo, K, Collins, H, Quesada-Ocampo, L (2019) Evaluation of 
cultivars for control of downy mildew on squash, Kinston 2018. Plant Dis Manag 
Reports 13: V124 
 
Ali, S., Gautier, A., Leconte, M., Enjalbert, J., de Vallavieille-Pope, C. (2011) A rapid 
genotyping method for an obligate fungal pathogen, Puccinia striiformis f.sp. tritici, 
based on DNA extraction from infected leaf and multiplex PCR genotyping. BMC Res 
Notes 4: 240 
 
Billard, A., Laval, V., Fillinger, S., Leroux, P., Lachaise, H., Beffa, R., and Debieu, D. 
(2012) The allele-specific probe and primer amplification assay, a new real-time PCR 
method for fine quantification of single-nucleotide polymorphisms in pooled DNA. Appl 
Environ Microbiol 78: 1063–1068 
 
Blum, M., Waldner, M., Olaya, G., Cohen, Y., Gisi, U., and Sierotzki, H. (2011) 
Resistance mechanism to carboxylic acid amide fungicides in the cucurbit downy 
mildew pathogen Pseudoperonospora cubensis. Pest Manag Sci 67: 1211–1214 
 
Bolger, A.M., Lohse, M., Usadel, B. (2014) Trimmomatic: a flexible trimmer for Illumina 
sequence data. Bioinformatics 30: 2114–2120 
 
Bourret, T.B., Choudhury, R.A., Mehl, H.K., Blomquist, C.L., McRoberts, N., Rizzo, D.M. 
(2018) Multiple origins of downy mildews and mito-nuclear discordance within the 
paraphyletic genus Phytophthora. PLoS One 13: e0192502 
 
Brown, A.H.D., Feldman, M.W., Nevo, E. (1980) Multilocus structure of natural 
populations of Hordeum spontaneum. Genetics 96: 523-536. 
 
Brzozowski, L., Holdsworth, W., Mazourek, M. (2016) ‘DMR-NY401’: A new downy 
mildew–resistant slicing cucumber. HortScience 51: 1294–1296 
 

 

26 

 

 

Burkhardt, A., Buchanan, A., Cumbie, J.S, Savory, E.A., Chang, J.H., Day, B. (2014) 
Alternative splicing in the obligate biotrophic oomycete pathogen Pseudoperonospora 
cubensis. Mol Plant-Microbe Interact 28: 298–309 
 
Burkhardt, A., Day, B. (2016) Transcriptome and small RNAome dynamics during a 
resistant and susceptible interaction between cucumber and downy mildew. The Plant 
Genome: 9:1-19. 
 
Bustin, S. A., Benes, V., Garson, J.A., Hellemans, J., Huggett, J., Kubista, M., Mueller, 
R., Nolan, T., Pfaffl, M.W., Shipley, G.L., Vandesompele, J., and Wittwer, C. T (2009) 
The MIQE guidelines: minimum information for publication of quantitative real-time PCR 
experiments. Clin Chem 55: 611–622 
 
Call, A.D., Criswell, A.D., Wehner, T.C., Ando, K., Grumet, R. (2012a) Resistance of 
cucumber cultivars to a new strain of cucurbit downy mildew. HortScience 47: 171–178 
 
Call, A.D., Criswell, A.D., Wehner, T.C., Klosinska, U., Kozik, E.U. (2012b) Screening 
cucumber for resistance to downy mildew caused by Pseudoperonospora cubensis 
(Berk. and Curt.) Rostov. Crop Sci 52: 577–592 
 
Carisse, O., McNealis, V., Kriss, A. (2017) Association between weather variables, 
airborne inoculum concentration, and raspberry fruit rot caused by Botrytis cinerea. 
Phytopathology 108: 70–82 
 
Carisse, O., Tremblay, D.M., Lévesque, C.A., Gindro, K., Ward, P., and Houde, A. 
(2009) Development of a Taqman real-time PCR assay for quantification of airborne 
conidia of Botrytis squamosa and management of Botrytis leaf blight of onion. 
Phytopathology 99: 1273–1280 
 
Carleson NC, Fieland VJ, Scagel CF, Weiland JE, Grünwald NJ (2019) Population 
structure of Phytophthora plurivora on Rhododendron in Oregon nurseries. Plant Dis 
103: 1923–1930 
 
Chee, H.Y., Nelson, M.E., Grov,e G.G., Eastwell, K.C., Kenny, S.T., Klein, R.E. (2006) 
Population biology of Pseudoperonospora humuli in Oregon and Washington. Plant Dis 
90: 1283–1286 
 
Choi, Y-J., Hong, S-B., Shin, H-D. (2005) A re-consideration of Pseudoperonospora 
cubensis and P. humuli based on molecular and morphological data. Mycol Res 109: 
841–848 
 

Choi, Y-J., Thines, M. (2015) Host jumps and radiation, not co-divergence drives 

diversification of obligate pathogens. a case study in downy mildews and Asteraceae. 
PLoS One 10: e0133655 
 

 

27 

 

 

Choudhury, R.A., Koike, S.T., Fox, A.D., Anchieta, A., Subbarao, K.V., Klosterman, 
S.J., McRoberts, N. (2016b) Spatiotemporal patterns in the airborne dispersal of 
spinach downy mildew. Phytopathology 107: 50–58 
 
Clark, M.J., Chen, R, Lam, H.Y.K., Karczewski, K.J., Chen, R., Euskirchen, G., Butte, 
A.J., Snyder, M. (2011) Performance comparison of exome DNA sequencing 
technologies. Nat Biotechnol 29: 908–914 
Cohen, Y., Van den Langenberg, K.M., Wehner, T.C, Ojiambo, P.S., Hausbeck, M, 
Quesada-Ocampo, L.M., Lebeda, A., Sierotzki, H., and Gisi, U. (2015) Resurgence 
of Pseudoperonospora cubensis: the causal agent of cucurbit downy mildew. 
Phytopathology 105: 998–1012 
 
Cohen, Y., Meron, I., Mor, N., Zuriel, S. (2003) A new pathotype of Pseudoperonospora 
cubensis causing downy mildew in cucurbits in Israel. Phytoparasitica 31: 458–466 
 
Cohen, Y., and Rotem, J. (1971) Dispersal and viability of sporangia of 
Pseudoperonospora cubensis. Trans Br Mycol Soc 57: 67–74 
 
Cohen, Y., Rubin, A.E., Galperin, M. (2011) Formation and Infectivity of oospores of 
Pseudoperonospora cubensis, the causal agent of downy mildew in cucurbits. Plant Dis 
95: 874 
 
Cohen, Y., Rubin, A.E., Galperin, M, Ploch, S., Runge, F., Thines, M. (2014) Seed 
transmission of Pseudoperonospora cubensis. PLoS One 9: e109766 
 
Cohen, Y., Samoucha, Y. (1983) Cross-resistance to four systemic fungicides in 
metalaxyl-resistant strains of Phytophthora infestans and Pseudoperonospora cubensis. 
Plant Dis 68: 137-139. 
 
Coley-Smith, J.R. (1962) Overwintering of hop downy mildew Pseudoperonospora 
humuli (Miy. and Tak.) Wilson. Ann Appl Biol 50: 235–243 
 
Cruz-Dávalos, D.I., Llamas, B., Gaunitz, C., Fages, A., Gamba, C., Soubrier, J., 
Librado, P., Seguin-Orlando, A., Pruvost, M., Alfarhan, A.H., et al (2017) Experimental 
conditions improving in-solution target enrichment for ancient DNA. Mol Ecol Resour 17: 
508–522 
 
Cui, C., Herlihy, J.H., Bombarely, A., McDowell, J.M., Haak, D.C. (2019) Draft assembly 
of Phytophthora capsici from long-read sequencing uncovers complexity. Mol Plant-
Microbe Interact 32: 1559–1563 
 
Desjardins, C.A., Giamberardino, C., Sykes, S.M., Yu, C-H., Tenor, J.L., Chen, Y., 
Yang, T., Jones, A.M., Sun, S., Haverkamp, M.R., et al (2017) Population genomics and 
the evolution of virulence in the fungal pathogen Cryptococcus neoformans. Genome 
Res 27: 1207-1219 
 

 

28 

 

 

Dhar, N., Mamo, B.E., Subbarao, K., Koike, S., Fox, A., Anchieta, A., and Klosterman, 
S.J. (2019) Measurements of aerial spore load by qPCR facilitates lettuce downy 
mildew risk advisement. Plant Dis 104: 82–93 
 
Dung, J.K.S, Scott, J.C., Cheng, Q., Alderman, S.C., Kaur, N., Walenta, D.L., Frost, 
K.E., and Hamm, P.B. (2018) Detection and quantification of airborne Claviceps 
purpurea sensu lato ascospores from Hirst-type spore traps using real-time quantitative 
PCR. Plant Dis 102: 2487–2493 
Dutta, B. (2020) Evaluation of individual fungicides for downy mildew control in 
cucumber in Tift County, Georgia, 2019. Plant Dis Manag Rep Report 14: V086 
 
Ernest, E.G., Kee, W.E., and Everts, K.L. (2005) Evaluation of fungicides for control of 
downy mildew on pickling cucumber, Trial 2, 2005. Plant Dis Manag Rep 2: V066 
 
Evenhuis, A., Verdam, B., Zadoks, J.C. (2003) Splash dispersal of conidia of 
Mycocentrospora acerina in the field. Plant Pathol 46: 459–469 
 
Faircloth, B.C., Branstetter, M.G., White, N.D., Brady, S.G. (2015) Target enrichment of 
ultraconserved elements from arthropods provides a genomic perspective on 
relationships among Hymenoptera. Mol Ecol Resour 15: 489–501 
 
Fletcher, K., Gil, J., Bertier, L.D., Kenefick, A., Wood, K.J., Zhang, L., Reyes-Chin-Wo, 
S., Cavanaugh, K., Tsuchida, C., Wong, J., et al (2019) Genomic signatures of 
heterokaryosis in the oomycete pathogen Bremia lactucae. Nat Commun 10: 2645 
 
Fungicide Resistance Action Committee (FRAC) (2018) ode List 2018: Fungicides 
sorted by mode of action (including FRAC Code numbering).  
 
Gent, D.H., Adair, N., Knaus, B.J., Grünwald, N.J. (2019) Genotyping-by-sequencing 
reveals fine-scale population differentiation in populations of Pseudoperonospora 
humuli. Phytopathology. doi: 10.1094/PHYTO-12-18-0485-R 
 
Gent, D.H., Nelson, M.E., Farnsworth, J.L., Grove, G.G. (2009) PCR detection 
of Pseudoperonospora humuli in air samples from hop yards. Plant Pathol 58: 1081–
1091 
 
Gent, D.H., Ocamb, C.M. (2009) Predicting Infection Risk of Hop by Pseudoperonspora 
humuli. Phytopathology 99: 1190–1198 
 
Gent DH, Ocamb CM, Farnsworth JL (2010) Forecasting and Management of hop 
downy mildew. Plant Dis 94: 425–431 
 
Gevens, A.J., and Hausbeck, M.K. (2005) Control of downy mildew of cucumbers with 
fungicides, 2005. Plant Dis Manag Rep 2: V066 
 

 

29 

 

 

Gisi U, Sierotzki H (2015) Oomycete fungicides: phenylamides, quinone outside 
inhibitors, and carboxylic acid amide. In H Ishii, DW Hollomon (eds) Fungicide 
resistance in plant pathogens: Principles and a guide to practical management. Springer 
Japan, Tokyo, pp 145–174 
 
Gisi, U., Sierotzki, H. (2008) Fungicide modes of action and resistance in downy 
mildews. Eur J Plant Pathol 122: 157–167 
 
Gnirke, A., Melnikov, A., Maguire, J., Rogov, P., LeProust., E.M., Brockman, W., 
Fennell, T., Giannoukos, G., Fisher, S., Russ, C., et al (2009) Solution hybrid selection 
with ultra-long oligonucleotides for massively parallel targeted sequencing. Nat 
Biotechnol 27: 182–189 
 
Goldenhar, K.E., Hausbeck, M.K. (2019) Fungicides for control of downy mildew on 
pickling cucumber in Michigan. Plant Heal Prog 20: 165–169 
 
Goss, E.M., Press, C.M., Grünwald, N.J. (2013) Evolution of RXLR-class effectors in the 
oomycete plant pathogen Phytophthora ramorum. PLoS One 8: e79347 
 
Goudet, J. (2005) hierfstat, a package for r to compute and test hierarchical F-statistics. 
Mol Ecol Notes 5: 184–186 
 
Granke, L.L., and Hausbeck, M.K. (2011) Dynamics of Pseudoperonospora 
cubensis sporangia in commercial cucurbit fields in Michigan. Plant Dis 95: 1392–1400 
 
Granke, L.L., Morrice, J.J., and Hausbeck, M.K. (2013) Relationships between 
airborne Pseudoperonospora cubensis sporangia, environmental conditions, and 
cucumber downy mildew severity. Plant Dis 98: 674–681 
 
Grünwald, N.J., Everhart, S.E., Knaus, B.J., Kamvar, Z.N. (2017) Best practices for 
population genetic analyses. Phytopathology 107: 1000–1010 
 
Grünwald, N.J., McDonald, B.A., Milgroom, M.G. (2016) Population genomics of fungal 
and oomycete pathogens. Annu Rev Phytopathol 54: 323–346 
 
Gugino, B.K., Grove, T.L. (2020) Evaluation of select fungicides and fungicide programs 
for managing downy mildew of cucumber, 2019. Plant Dis Manag Rep Report 14: V005 
 
Gugino, B.K., and Grove, T.L. (2016) Evaluation of selected fungicides and fungicide 
programs for managing downy mildew of cucumber, 2015. Plant Dis. Manag. Rep. 
10:V072: 
 
Hartl, D., Clark, A. (1997) Principles of population genetics. Sinauer Associates, Inc, 
Sunderland, Massachusetts 
 

 

30 

 

 

Hausbeck, M. K., and Linderman, S.D. (2014) Evaluation of fungicides for control of 
downy mildew of cucumber, 2013. Plant Dis Manag Rep 8: V304 
 
Hausbeck, M., Brisco-McCann, E., Goldenhar, K., and Cook, A. (2017) Evaluation of 
fungicide programs for control of downy mildew of cucumber, 2016. Plant Dis Manag 
Reports 11: V063 
Hausbeck, M., Goldenhar, K. (2017) Downy mildew recommendations for cucumbers.  
MSU Ext., 
https://www.canr.msu.edu/news/downy_mildew_recommendations_for_cucumbers 
 
Hausbeck, M.K., and Cortright, B. (2010) Evaluation of fungicides and application 
methods for control of downy mildew of cucumber, 2010. Plant Dis. Manag. Rep. V6: 
 
Hausbeck, M.K., Perla, D.E., and Cook, A.J. (2019) Evaluation of single fungicide 
products for control of downy mildew of cucumber, 2018. Plant Dis Manag Reports 139: 
V13 
 
Holdsworth, W.L., Summers, C.F., Glos, M., Smart, C.D., Mazourek, M. (2014) 
Development of downy mildew-resistant cucumbers for late-season production in the 
Northeastern United States. HortScience horts 49: 10–17 
 
Hollomon, B. (2007) Fungicide resistance in crop pathogens: How can it be managed?. 
2nd ed. FRAC, Brussels 
 
Holmes, G.J., Ojiambo, P.S., Hausbeck, M.K., Quesada-Ocampo, L., and Keinath, A.P. 
(2014) Resurgence of cucurbit downy mildew in the United States: a watershed event 
for research and extension. Plant Dis 99: 428–441 
 
Jackson, S.L., Bayliss, K.L. (2011) Spore traps need improvement to fulfil plant 
biosecurity requirements. Plant Pathol 60: 801–810 
 
Jenkyn, J. (1974) A comparison of seasonal changes in deposition of spores of 
Erysiphe graminis on different trapping surfaces. Ann Appl Biol 76: 257–267 
 
Johnson, D.A, Alldredge, J.R, Allen, J.R, Allwine, R. (1991) Spatial pattern of downy 
mildew in hop yards during severe and mild disease epidemics. Phytopathology 81: 
1369–1374 
 
Jombart, T., and Ahmed, I. (2011) adegenet 1.3-1: new tools for the analysis of 
genome-wide SNP data. Bioinformatics 27: 3070–3071 
 
Jones, L., Riaz, S., Morales-Cruz, A., Amrine, K.C.H., McGuire, B., Gubler, W.D., 
Walker, M.A., Cantu, D. (2014) Adaptive genomic structural variation in the grape 
powdery mildew pathogen, Erysiphe necator. BMC Genomics 15: 1081 
 

 

31 

 

 

Kamvar, Z.N., Brooks, J.C., Grünwald, N.J. (2015) Novel R tools for analysis of 
genome-wide population genetic data with emphasis on clonality. Front Genet 6: 208 
 
Kamvar, Z.N., Tabima, J.F., Grünwald, N.J. (2014) Poppr: an R package for genetic 
analysis of populations with clonal, partially clonal, and/or sexual reproduction. PeerJ 2: 
e281 
 
Kanetis, L., Holmes, G.J, and Ojiambo, P.S. (2010) Survival of Pseudoperonospora 
cubensis sporangia exposed to solar radiation. Plant Pathol 59: 313–323 
 
Kanetis, L., Vasiliou, A., Neophytou, G., Samouel, S., Tsaltas, D. (2013) First report of 
downy mildew caused by Peronospora belbahrii on sweet basil (Ocimum basilicum) in 
Cyprus. Plant Dis 98: 283 
 
Keinath, A. P. (2015) Utility of a cucumber plant bioassay to assess fungicide efficacy 
against Pseudoperonospora cubensis. Plant Dis 100: 490–499 
 
Keinath, A.P., Wintermantel WM, Zitter TA, eds (2017) Compendium of cucurbit 
diseases and pests, 2nd ed. St. Paul, Minnesota, USA 
 
Kitner, M., Lebeda, A., Sharma, R., Runge, F., Dvořák, P., Tahir, A., Choi, Y-J., 
Sedláková, B., Thines, M. (2015) Coincidence of virulence shifts and population genetic 
changes of Pseudoperonospora cubensis in the Czech Republic. Plant Pathol 64: 
1461–1470 
 
Klosterman, S.J., Anchieta, A., McRoberts, N., Koike, S.T., Subbarao, K. V., Voglmayr, 
H., Choi, Y-J., Thines, M., and Martin, F.N. (2014) Coupling spore traps and quantitative 
PCR assays for detection of the downy mildew pathogens of spinach (Peronospora 
effusa) and beet (P. schachtii). Phytopathology 104: 1349–1359 
 
Knaus, B.J., Grünwald, N.J. (2017) vcfr: a package to manipulate and visualize variant 
call format data in R. Mol Ecol Resour 17: 44–53 
 
Kozarewa, I., Armisen, J., Gardner, A.F., Slatko, B.E., Hendrickson, C.L. (2015) 
Overview of target enrichment strategies. Curr Protoc Mol Biol 112: 7.21.1-7.21.23 
 
Kunjeti, S.G., Anchieta, A., Martin ,F.N., Choi, Y-J., Thines, M., Michelmore, R.W., 
Koike, S.T., Tsuchida, C., Mahaffee, W., Subbarao, K. V., Klosterman, S. J. (2016) 
Detection and quantification of Bremia lactucae by spore trapping and quantitative PCR. 
Phytopathology 106: 1426–1437 
 
Lane, C.R., Beales, P.A., O’Neill, T.M., McPherson, G.M., Finlay, A.R., David, J., 
Constantinescu, O., Henricot, B. (2005) First report of Impatiens downy mildew 
(Plasmopara obducens) in the UK. Plant Pathol 54: 243 
 

 

32 

 

 

Langston, D., Sanders, F.H. (2013) Evaluation of fungicides for control of downy mildew 
on cucumber in Georgia I, 2012. Plant Dis. Manag. Rep. 7: V108. 
 
Lebeda, A., Urban, J. (2010) Screening for resistance to cucurbit downy mildew 
(Pseudoperonospora cubensis). MASS Screen. Tech. Sel. Crop. Resist. TO Dis. 
International Atomic Energy Agency, Viena, pp 284–295 
 
Li, C-S., Lin, Y-C. (1999) Sampling performance of impactors for fungal spores and 
yeast cells. Aerosol Sci Technol 31: 226–230 
 
Li, H., Durbin, R. (2009) Fast and accurate short read alignment with Burrows-Wheeler 
transform. Bioinformatics. doi: 10.1093/bioinformatics/btp324 
 
Li, H., Handsaker, B., Wysoker, A., Fennell, T., Ruan J., Homer, N., Marth, G., 
Abecasis, G., Durbin, R., Subgroup 1000 Genome Project Data Processing (2009) The 
sequence alignment/map format and SAMtools. Bioinformatics 25: 2078–2079 
 
Lim, H.C., Braun, M.J. (2016) High-throughput SNP genotyping of historical and modern 
samples of five bird species via sequence capture of ultraconserved elements. Mol Ecol 
Resour 16: 1204–1223 
 
Lizotte, E., and Miles, T.D. (2020) Hop disease management in Michigan for 2020. MSU 
Ext., https://www.canr.msu.edu/news/managing-hop-downy-mildew-in-michigan 
 
Ma, X., Shao, Y., Tian, L., Flasch, D.A., Mulder, H.L., Edmonson, M.N., Liu, Y., Chen, 
X., Newman, S., Nakitandwe, J., et al (2019) Analysis of error profiles in deep next-
generation sequencing data. Genome Biol 20: 50 
 
McCormack JE, Tsai WLE, Faircloth BC (2016) Sequence capture of ultraconserved 
elements from bird museum specimens. Mol Ecol Resour 16: 1189–1203 
 
Michigan Department of Agriculture & Rural Development (2018) Michigan Agriculture 
Facts & Figures. 
https://www.michigan.gov/documents/mdard/MI_Ag_Facts__Figures_474011_7.pdf 
 
Milgroom, M.G. (2015) Population biology of plant pathogens: genetics, ecology, and 
evolution. The American Phytopathological Society, St. Paul, Minnesota 
 
Milgroom, M.G. (1996) Recombination and the multilocus structure of fungal 
populations. Annu Rev Phytopathol 34: 457–477 
 
Miller SA, Vargas A, Amrhein J (2020) Bioassay for sensitivity of Pseudoperonospora 
cubensis to fungicides in OH, 2019. Plant Dis Manag Rep Report 14: V024 
 

 

33 

 

 

Mitchell, M.N., Ocamb, C.M., Grünwald, N.J., Mancino, L.E., and Gent, D.H. (2011) 
Genetic and pathogenic relatedness of Pseudoperonospora cubensis and P. humuli. 
Phytopathology 101: 805–818 
 
Naegele, R. P., Quesada-Ocampo, L. M. , Kurjan, J. D., Saude, C., and Hausbeck, M. 
K. (2016) Regional and temporal population structure of Pseudoperonospora 
cubensis in Michigan and Ontario. Phytopathology 106: 372–379 
 
Nei, M (1972) Genetic distance between populations. Am Nat 106: 283–292 
Neufeld, K.N., Keinath, A.P., Ojiambo, P.S. (2017) Evaluation of a model for predicting 
the infection risk of squash and cantaloupe by Pseudoperonospora cubensis. Plant Dis 
102: 855–862 
 
Ojiambo, P.S., Gent, D.H., Quesada-Ocampo, L.M., Hausbeck, M.K., and Holmes, G.J. 
(2015) Epidemiology and population biology of Pseudoperonospora cubensis: a model 
system for management of downy mildews. Annu Rev Phytopathol 53: 223–246 
 
Ojiambo, P.S., and Holmes, G.J. (2010) Spatiotemporal spread of cucurbit downy 
mildew in the Eastern United States. Phytopathology 101: 451–461 
 
Ojiambo, P.S., Holmes, G.J., Britton, W., Babadoost, M., Bost, S.C., Boyles, R., Brooks, 
M., Damicone, J., Draper, M.A., Egel, D.S., et al (2011) Cucurbit Downy Mildew 
ipmPIPE: A next generation web-based interactive tool for disease management and 
extension outreach. Plant Heal Prog 12: 26 
 
Oyola, S.O., Gu, Y., Manske, M., Otto, T.D., O’Brien, J., Alcock, D., MacInnis, B., 
Berriman, M., Newbold, C.I., Kwiatkowski, D.P, et al (2013) Efficient depletion of host 
DNA contamination in malaria clinical sequencing. J Clin Microbiol 51: 745 LP – 751 
 
Perea, C., De La Hoz, J.F., Cruz, D.F., Lobaton, J.D., Izquierdo, P., Quintero, J.C., 
Raatz, B., Duitama, J. (2016) Bioinformatic analysis of genotype by sequencing (GBS) 
data with NGSEP. BMC Genomics 17: 498 
 
Perla, D., Hausbeck, M.K., Cook, A. (2019) Evaluation of fungicide programs for control 
of downy mildew of cucumber, 2018. Plant Dis Manag Reports 147: V13 
 
Potapov, V., Ong, J.L. (2017) Examining sources of error in PCR by single-molecule 
sequencing. PLoS One 12: e0169774 
 
Quesada-Ocampo L.M., Granke, L.L., Olsen, J., Gutting, H.C., Runge, F., Thines, M., 
Lebeda, A., and Hausbeck, M.K. (2012) The genetic structure of Pseudoperonospora 
cubensis populations. Plant Dis 96: 1459–1470 
 
Quinlan, AR, Hall, IM (2010) BEDTools: a flexible suite of utilities for comparing 
genomic features. Bioinformatics 26: 841–842 
 

 

34 

 

 

Rahman, A., Góngora-Castillo, E., Bowman, M.J., Childs, K.L., Gent, D.H., Martin, F.N., 
Quesada-Ocampo, L.M. (2019) Genome sequencing and transcriptome analysis of the 
hop downy mildew pathogen Pseudoperonospora humuli reveal species-specific genes 
for molecular detection. Phytopathology 109: 1354–1366 
 
Reuveni, M., Eyal, H., Cohen, Y. (1980) Development of resistance to metalaxyl 
in Pseudoperonospora cubensis. Plant Dis 64: 1108–1109 
 
Rivera, Y., Salgado-Salazar, C., Gulya, T.J., Crouch, J.A. (2016) Newly emerged 
populations of Plasmopara halstedii infecting rudbeckia exhibit unique genotypic profiles 
and are distinct from sunflower-infecting strains. Phytopathology 106: 752–761 
 
Rogers, S.L., Atkins, S.D., and West, J.S. (2009) Detection and quantification of 
airborne inoculum of Sclerotinia sclerotiorum using quantitative PCR. Plant Pathol 58: 
324–331 
 
Royle, D.J. (1973) Quantitative relationships between infection by the hop downy 
mildew pathogen, Pseudoperonospora hutnuli, and weather and inoculum factors. Ann 
Appl Biol 73: 19–30 
 
Runge, F., and Thines, M. (2011) Reevaluation of host specificity of the closely related 
species Pseudoperonospora humuli and P. cubensis. Plant Dis 96: 55–61 
 
Savory, E.A., Adhikari, B.N., Hamilton, J.P., Vaillancourt, B., Buell, C.R., Day, B. 
(2012a) mRNA-Seq analysis of the Pseudoperonospora cubensis transcriptome during 
cucumber (Cucumis sativus L.) infection. PLoS One 7: e35796 
 
Savory, E.A., Granke, L.L., Quesada-Ocampo, L.M., Varbanova, M., Hausbeck, M.K., 
Day, B. (2011) The cucurbit downy mildew pathogen Pseudoperonospora cubensis. Mol 
Plant Pathol 12: 217–26 
 
Savory, E.A., Zou, C., Adhikari, B.N., Hamilton, J.P., Buell, C.R., Shiu, S-H., Day, B. 
(2012b) Alternative splicing of a multi-drug transporter from Pseudoperonospora 
cubensis generates an RXLR effector protein that elicits a rapid cell death. PLoS One 7: 
e34701 
 
Sharma, R., Xia, X., Cano, L.M., Evangelisti, E., Kemen, E., Judelson, H., Oome, S., 
Sambles, C., van den Hoogen, D.J., Kitner, M., et al (2015) Genome analyses of the 
sunflower pathogen Plasmopara halstedii provide insights into effector evolution in 
downy mildews and Phytophthora. BMC Genomics 16: 741 
 
Sitterly, W.R. (1972) Breeding for disease resistance in cucurbits. Annu Rev 
Phytopathol 10: 471–490 
 

 

35 

 

 

Starrett, J., Derkarabetian, S., Hedin, M., Bryson, Jr. R.W., McCormack, J.E., Faircloth, 
B.C. (2017) High phylogenetic utility of an ultraconserved element probe set designed 
for Arachnida. Mol Ecol Resour 17: 812–823 
 
Stassen, J., Seidl, M., Vergeer, P., Nijman, I., Snel, B., Cuppen, E., Van den 
ackerveken, G. (2012) Effector identification in the lettuce downy mildew Bremia 
lactucae by massively parallel transcriptome sequencing. Mol Plant Pathol 13: 719–731 
 
Summers, C., Adair, N., Gent, D., Mcgrath, M., and Smart, C. (2015a) 
Pseudoperonospora cubensis and P. humuli detection using species-specific probes 
and high definition melt curve analysis. Can J Plant Pathol 37: 150605144651006 
 
Summers C.F., Gulliford C.M., Carlson C.H., Lillis J.A., Carlson M.O., Cadle-Davidson 
L., Gent, D.H., Smart, C.D. (2015b) Identification of genetic variation between obligate 
plant pathogens Pseudoperonospora cubensis and P. humuli using RNA sequencing 
and genotyping-by-sequencing. PLoS One 10: e0143665 
 
Tabima, J.F., Coffey, M.D., Zazada, I.A., Grünwald, N.J. (2018) Populations 
of Phytophthora rubi show little differentiation and high rates of migration among states 
in the western United States. Mol Plant-Microbe Interact 31: 614–622 
 
Thines, M. (2014) Phylogeny and evolution of plant pathogenic oomycetes—a global 
overview. Eur J Plant Pathol 138: 431–447 
 
Thines, M., Choi, Y-J. (2015) Evolution, Diversity, and Taxonomy of the 
Peronosporaceae, with Focus on the Genus Peronospora. Phytopathology 106: 6–18 
 
Thomas, A., Carbone, I., Ojiambo, P. (2013) Comparative genomic analysis of 
Pseudoperonospora cubensis to elucidate the genetic basis of host specialization. 
Phytopathology 103: S2.145 
 
Thomas, A., Carbone, I., Choe, K., Quesada-Ocampo, L.M., and Ojiambo, P.S. (2017a) 
Resurgence of cucurbit downy mildew in the United States: Insights from comparative 
genomic analysis of Pseudoperonospora cubensis. Ecol Evol 7: 6231–6246 
 
Thomas, A., Carbone, I., Cohen, Y., Ojiambo, P.S. (2017b) Occurrence and Distribution 
of Mating Types of Pseudoperonospora cubensis in the United States. Phytopathology 
107: 313–321 
 
Thomas, A., Carbone, I., Lebeda, A., Ojiambo, P.S. (2017c) Virulence structure within 
populations of Pseudoperonospora cubensis in the United States. Phytopathology 107: 
777–785 
 
Thomas, A., Neufeld, K.N., Seebold, K.W., Braun, C.A., Schwarz, M.R., and Ojiambo, 
P.S. (2018) Resistance to fluopicolide and propamocarb and baseline sensitivity to 

 

36 

 

 

ethaboxam among isolates of Pseudoperonospora cubensis from the Eastern United 
States. Plant Dis 102: 1619–1626 
 
Thornton, A.C, Adams, M.L., and Holmes, G.J. (2006) Evaluation of fungicides for 
control of downy mildew of cucumber I, 2006. Plant Dis Manag Rep 1: V122 
 
McGrath, M (2004) Diseases of cucurbits and their management BT - Diseases of fruits 
and vegetables volume I: diagnosis and management. In SAMH Naqvi (ed) Springer 
Netherlands, Dordrecht, pp 455–510 
Urban, J., Lebeda, A. (2006) Fungicide resistance in cucurbit downy mildew – 
methodological, biological and population aspects. Ann Appl Biol 149: 63–75 
 
USDA (2020) United States Department of Agriculture. 2019. Vegetables Annual 
Summary. https://usda-library-cornell-
edu.proxy1.cl.msu.edu/concern/publications/02870v86p?locale=en 
 
Vleeshouwers, V.G., Rietman, H, Krenek, P., Champouret, N., Young, C., Oh, S-K., 
Wang, M., Bouwmeester, K., Vosman, B., Visser, R.G.F., et al (2008) Effector genomics 
accelerates discovery and functional profiling of potato disease resistance and 
Phytophthora Infestans avirulence genes. PLoS One 3: e2875 
 
Wallace, E., Adams, M., Ivors, K., Ojiambo, P.S., Quesada-Ocampo, L.M. (2014) First 
report of Pseudoperonospora cubensis causing downy mildew on Momordica balsamina 
and M. charantia in North Carolina. Plant Dis 98: 1279 
 
Wallace, E., Adams, M., Quesada-Ocampo, L.M. (2015) First report of downy mildew on 
buffalo gourd (Cucurbita foetidissima) caused by Pseudoperonospora cubensis in North 
Carolina. Plant Dis 99: 1861 
 
Wallace, E., Quesada-Ocampo, L. (2016) Pseudoperonospora cubensis on commercial 
and non-commercial cucurbits in North Carolina: Population structure determined by 
simple sequence repeats. Phythopathology 106: (Suppl) S412 
 
Wallace, E.C., D’Arcangelo, K.N., Quesada-Ocampo, L.M. (2020) Population analyses 
reveal two host-adapted clades of Pseudoperonospora cubensis, the causal agent of 
cucurbit downy mildew, on commercial and wild cucurbits. Phytopathology. doi: 
10.1094/PHYTO-01-20-0009-R 
 
Wallace, E.C., Quesada-Ocampo, L.M. (2017) Analysis of microsatellites from the 
transcriptome of downy mildew pathogens and their application for characterization of 
Pseudoperonospora populations. PeerJ 5: e3266 
 
Withers, S., Gongora-Castillo, E., Gent, D., Thomas, A., Ojiambo, P.S., and Quesada-
Ocampo, L.M. (2016) Using Next-Generation sequencing to develop molecular 
diagnostics for Pseudoperonospora cubensis, the cucurbit downy mildew pathogen. 
Phytopathology 106: 1105–1116 

 

37 

 

 

 
Wong, F.P, and Wilcox, W.F (2001) Comparative physical modes of action of 
azoxystrobin, mancozeb, and metalaxyl gainst Plasmopara viticola (grapevine downy 
mildew). Plant Dis 85: 649–656 
 
Wu, E-J., Yang, L-N., Zhu, W., Chen, X-M., Shang, L-P., Zhan, J. (2016) Diverse 
mechanisms shape the evolution of virulence factors in the potato late blight pathogen 
Phytophthora infestans sampled from China. Sci Rep 6: V26182 
 
Zhang, Y., Pu, Z., Qin, Z., Zhou, X., Liu, D., Dai, L., Wang, W. (2012) A study on the 
overwintering of cucumber downy mildew oospores in China. J Phytopathol 160: 469–
474 
 
Zhu, S.S., Liu, X.L., Wang, Y., Wu, X.H., Liu, P.F., Li, J.Q., Yuan, S.K., and Si, N.G. 
(2007) Resistance of Pseudoperonospora cubensis to flumorph on cucumber in plastic 
houses. Plant Pathol 56: 967–975 
 
 
 
 
 
 
 
 
 
 
 

 

 

 

 

 

 

 
 
 
 

 

38 

 

 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CHAPTER II: DETECTION OF AIRBORNE SPORANGIA OF 

PSEUDOPERONOSPORA CUBENSIS AND P. HUMULI IN MICHIGAN USING 

BURKARD SPORE TRAPS COUPLED TO QPCR 

39 

 

 

ABSTRACT  
 

Cucurbit downy mildew (CDM), caused by the oomycete pathogen 

Pseudoperonospora cubensis, is a devastating foliar disease on cucumber resulting in 

reduced yields. In 2004, the pathogen re-emerged in the U.S., infecting historically 

resistant cucumber cultivars and requiring the adoption of an intensive fungicide 

program. The pathogen cannot overwinter in Michigan fields but due to an influx of 

airborne sporangia cucurbit downy mildew occurs annually. In Michigan, spore traps are 

used to monitor the presence of airborne P. cubensis sporangia in cucumber growing 

regions to guide the initiation of a fungicide program. However, Pseudoperonospora 

humuli sporangia, the causal agent of downy mildew on hop, are morphologically 

indistinguishable from P. cubensis sporangia. This morphological similarity reduces the 

ability to accurately detect P. cubensis from spore trap samples when examined with 

the aid of light microscopy. To improve P. cubensis detection, we adapted a qPCR-

based assay to allow the differentiation between P. cubensis and P. humuli on Burkard 

spore trap samples collected in the field. Specifically, we evaluated the specificity and 

sensitivity of P. cubensis detection on Burkard spore trap tapes using a morphological 

based and qPCR-based identification assay and determined whether sporangia of P. 

cubensis and P. humuli on Burkard samples could be distinguished using qPCR. We 

found that the qPCR assay was able to detect a single sporangium of each species on 

spore trap samples collected in the field with Cq values below 35.5. The qPCR assay 

also allowed the detection of P. cubensis and P. humuli in samples containing 

sporangia from both species. However, the number of sporangia quantified using light 

microscopy explained only 54% and 10% of the variation in the Cq values of P. cubensis 

 

40 

 

 

and P. humuli, respectively, suggesting a limited capacity of the qPCR assay for the 

absolute quantification of sporangia in field samples. After two years of monitoring using 

Burkard spore traps coupled with the qPCR in cucumber fields, P. humuli sporangia 

were detected more frequently than P. cubensis early in the growing season (May and 

June). P. cubensis sporangia were detected approximately 5 -10 days before cucurbit 

downy mildew symptoms were first observed in cucumber fields during both years. This 

research describes an improved sporangial detection system that is key for the 

monitoring and management of P. cubensis in Michigan.   

INTRODUCTION 
 

Pseudoperonospora cubensis (Berk. & M. A. Curtis) Rostovzev, the causal agent 

of cucurbit downy mildew (CDM), infects approximately 20 cucurbit genera including the 

economically important crops of cucumber, cantaloupe, squash, watermelon and 

pumpkin (Cohen et al., 2015; Lebeda and Cohen, 2011). CDM symptoms include 

angular, chlorotic lesions that coalesce and become necrotic, resulting in leaf blight and 

death; pathogen sporulation occurs on the abaxial side of the leaf (Salcedo et al., 2020). 

In cucumber, foliar blighting resulting from CDM can result in yield reduction (Hausbeck 

et al., 2019; Perla et al., 2019; Reuveni et al., 1980).  

In the U.S., Michigan is the largest producer of pickling cucumber and the 

second-largest producer of cucumber for the fresh market, with approximately 4.1 

million cwt of cucumbers sold in 2018 (USDA, 2020). For nearly 40 years, resistant 

cucumber cultivars had been used successfully to mitigate CDM (Brzozowski et al., 

2016). In 2004, a highly virulent strain of P. cubensis emerged in the U.S. overcoming 

this host resistance (Thomas et al., 2017) and since then fungicides have been relied on 

 

41 

 

 

for control (Blum et al., 2011; Holmes et al., 2014). However, fungicide-resistant P. 

cubensis isolates have presented crop protection challenges. Single-site fungicides 

including mefenoxam and azoxystrobin were ineffective when the pathogen emerged in 

2004 (Ernest et al., 2005; Gevens and Hausbeck, 2005; Thornton et al., 2006). Since 

that time, P. cubensis resistance to dimethomorph (Zhu et al., 2007) and 

mandipropamid (Hausbeck and Cortright, 2010; Blum et al., 2011) has been reported in 

the U.S. (Holmes et al., 2014; Keinath, 2015). Similarly, reduced efficacy of fluopicolide 

against CDM has been observed in field trials in Michigan (Hausbeck and Linderman, 

2014), Georgia (Langston and Sanders, 2013), and North Carolina (Adams and 

Quesada-Ocampo, 2014). While propamomcarb was effective against CDM for several 

years, since 2013 its efficacy appeared to be compromised in field trials in North 

Carolina (Keinath, 2015; Thomas et al., 2018), Pennsylvania (Gugino and Grove, 2016), 

and Michigan (Hausbeck and Linderman, 2014; Hausbeck et al., 2017).  

P. cubensis is an obligate pathogen and its survival depends on the availability of 

susceptible hosts (Cohen et al., 2015). The pathogen does not survive in regions that 

experience frost, instead its sporangia are dispersed to northern latitudes from 

overwintering sites (Ojiambo and Holmes, 2010; Quesada-Ocampo et al., 2012). 

Airborne sporangia concentrations influence CDM onset (Granke et al., 2013) and 

under conducive weather conditions, P. cubensis sporangia can spread rapidly within 

and between fields (Ojiambo et al, 2015). Airborne concentrations of P. cubensis 

sporangia in Michigan’s cucumber fields have been monitored using Burkard spore 

traps (Burkard Manufacturing Co Ltd, U.K.) with light microscopy used to identify and 

enumerate the pathogen sporangia based on morphology (Granke and Hausbeck, 

 

42 

 

 

2011; Granke et al., 2013). Pseudoperonospora humuli, the causal agent of hop downy 

mildew (HDM) is nearly identical to P. cubensis morphologically (Runge and Thines, 

2011) but rarely infects cucurbits in the U.S. (Mitchell et al., 2011). Approximately 400 

ha of hops are planted in Michigan (Michigan Department of Agriculture & Rural 

Development, 2018) and P. humuli is prevalent (Lizotte et al., 2020). Thus, relying on 

morphological identification, alone, to monitor airborne sporangial concentrations of P. 

cubensis could result in inaccurate estimations of the pathogen's presence and 

concentration. 

PCR-based methods have been used successfully to detect and quantify 

airborne plant pathogens such as Peronospora effusa (Klosterman et al., 2014), 

Peronospora schachtii (Klosterman et al., 2014) Claviceps purpurea (Dung et al., 2018) 

and P. humuli, infecting spinach, beet, grass-seed and hop (Gent et al., 2009), 

respectively. A qPCR assay was developed that differentiates between P. cubensis and 

P. humuli sporangia (Summers et al., 2015). This assay, with or without microscopic 

visualization of spore trap tapes, could accelerate the speed and accuracy of P. 

cubensis detection and inform the initiation of fungicide sprays. The objective of our 

study was to improve the detection of airborne concentrations of P. cubensis sporangia 

by adapting a qPCR-based assay (Summers et al., 2015) that distinguishes between P. 

cubensis and P. humuli using Burkard spore trap samples collected in the field.  

MATERIALS AND METHODS 
 

In vitro evaluations to assess the sensitivity of the qPCR-based assay were 

performed using isolates of P. cubensis (CDM23) and P. humuli (HDM19) obtained in 

2017 using methods similar to those described by Thomas et al. (2017). Briefly, 

 

43 

 

 

diseased tissue was placed in a moist chamber overnight to induce sporulation. 

Sporangia from a single cucumber leaf lesion or an infected hop basal shoots (spike) 

were suspended in 1 ml of distilled water and the resulting inoculum (1,000-10,000 

sporangia/ml) applied to the abaxial side of detached leaves of ‘Vlaspik’ cucumber or 

‘Centennial’ hop, respectively, contained in Petri dishes (100 X 15 mm). Inoculated 

leaves were then incubated in a growth chamber at 18°C under a 12/12-hr light/dark 

cycle. Seven to 10 days post inoculation, sporangia were gently rinsed from infected 

leaves using a Preval spray power unit (Preval, Chicago) filled with distilled water. A 

new set of leaves were inoculated with the resulting sporangia.  

Collection of sporangia and extraction of genomic DNA. P. cubensis 

(CDM23) and P. humuli sporangia (HDM19) were gently rinsed from host tissue into 

centrifuge tubes (50 ml) using a Preval spray power unit filled with distilled water. The 

sporangial suspension was concentrated by centrifugation (5424R centrifuge, 

Eppendorf) at 14,000 rpm for 5 min and homogenized in impact-resistant 2mL tubes 

(Lysing Matrix H, MP Biomedicals) using a TissueLyser II (Qiagen, Valencia, CA) for 

4min at 30 Hz. DNA was extracted using a NucleoSpin Plant II isolation kit (Macherey-

Nagel, Bethlehem, PA) following manufacturer’s instructions and DNA concentration 

was determined using the Qubit double-stranded DNA High Sensitivity Assay Kit (Life 

Technologies, Carlsbad, CA). 

Competitive qPCR internal control. A competitive internal positive control (IC) 

was designed in this study and incorporated into every qPCR reaction to monitor for the 

presence of PCR inhibitors in each sample. The IC consisted of a single-stranded linear 

synthetic DNA that utilizes the same primers of the target mitochondrial cox2 gene, and 

 

44 

 

 

an additional fluorogenic probe (ICprobeJ2: 

/5CYS/A+GCATTATT+GTTTAT+CATATATACA/3IABkFQ/) for amplification and 

detection. The sequence of the internal control (Table 2-1) showed no significant 

nucleotide identity to any known naturally occurring PCR-amplifiable nucleotide 

sequences reported in the NCBI database. 

qPCR protocol with purified DNA. All qPCR experiments were conducted 

using a modified version of the protocol described by Summers et al. (2015) in 

accordance with the Minimum Information for Publication of Quantitative Real-Time 

Experiments (MIQE) guidelines (Bustin et al., 2009). The Summers et al. (2015) assay 

was modified by changing the commercial master mix  IQ Supermix (Bio-rad, Hercules, 

CA) to the Prime-Time Gene Expression Master Mix (IDT, Skokie, IL). This new master 

mix reduced the variation between technical replicates and increased the amplification 

efficiency of qPCR reactions. Additionally, we also added an internal-positive control to 

identify any alterations in amplification efficiency in field air samples. qPCR reactions 

with a final volume of 20 µl were manually assembled in 96-well white plates (Bio-rad 

MLL9651) containing 10 µl of the Prime-Time Gene Expression Master Mix, 2 µl of 

template DNA, 600 nM of each primer (RT33F and RT182R), 500 nM of the LNA probe 

HUMprobeSNP105,  250 nM of the LNA probe CUBprobeSNP105, 250 nM of the LNA 

probe ICprobeJ2, and 7.5 x10-10 nM (0.75 aM) of our internal positive control (IC) (Table 

2-1). The IC was set to this concentration to obtain a Cq value of 29 without affecting the 

sensitivity or specificity of the other probes. Negative control reactions lacking the DNA 

template were included in every plate run. The qPCR protocol run on a CFX 96 Touch 

qPCR system (Bio-rad) and included an initial denaturation step at 95°C for 3 min 

 

45 

 

 

followed by 38 cycles of 95°C for 10 s and 65°C for 45 s. Two technical replicates of 

each sample were run and the average Cq and standard deviation were calculated 

using Bio-rad CFX Manager software (version 3.1) (Bio-rad 1845000).  

Sensitivity and specificity of qPCR with LNA probes. The Burkard spore 

traps use a vacuum pump to draw air (approximately 10 liters/min) into a collection 

chamber containing a reel, covered with a melinex tape; the reel was mounted on a 

clockwork mechanism (Hirst, 1952). The melinex tape (Burkard Manufacturing Co. Ltd., 

U.K.) was coated with an adhesive of petroleum jelly and paraffin (9:1 wt/wt) dissolved 

in sufficient toluene to provide adequate coverage of tape at the desired thickness 

(Granke et al., 2013).  

The sensitivity and specificity of the qPCR assay was evaluated using three 

different experiments; the first generated a standard curve using pure DNA of P. humuli  

and P. cubensis, the second used samples that contained a mixture of P. humuli and P. 

cubensis DNA and the third used samples of P. cubensis sporangia that also contained 

the melinex tape and the adhesive used on the Burkard spore trap tapes. The first 

procedure included ten-fold dilutions of genomic DNA from two independent DNA 

extractions from each pathogen isolate (CDM23 and HDM19) which were used to 

generate standard curves ranging from 10 to 100,000 fg. Three technical replicates of 

each sample dilution were tested using qPCR and the average Cq values with the 

standard deviation were calculated using the Bio-rad CFX Manager software (Bio-rad). 

Mean Cq values were plotted against the log10 of template DNA concentrations 

and used to generate standard curves. The second procedure included the evaluation 

of mixed-DNA samples to assess the specificity of the qPCR assay. An in vitro 

 

46 

 

 

assessment was used to determine whether the assay could detect P. cubensis and P. 

humuli in mixed samples containing DNA from both pathogens. Ten-fold dilutions of 

genomic DNA from each species were made from 100 to 100,000 fg and used as 

templates, both separately and mixed in varying concentrations (Table 2-2), for the 

qPCR assay described above. Three technical replicates of each concentration and 

mixture were run and the average Cq and standard deviations were calculated using 

Bio-rad CFX Manager software. The third procedure included five independent 

extractions from solutions containing 10, 100, or 1000 sporangia of P. cubensis 

prepared using a hemocytometer counting cell chamber. Sporangia were homogenized 

in impact-resistant 2 mL tubes using a TissueLyser (4 min at 30 Hz) and DNA was 

extracted using the NucleoSpin Plant II isolation kit (Macherey-Nagel, Bethlehem, PA). 

Subsequently, 2 µl of the extraction product were evaluated using qPCR. Finally, 9 x 48 

mm sections (representing a 24 h-sampling period) of melinex tape with adhesive were 

spiked with 10, 20, 50, 100 or 300 P. cubensis sporangia. DNA was extracted and 

evaluated using qPCR as previously described. Three technical replicates of an 

average of four independent extractions of each sample dilution were tested and the 

average Cq values with standard deviation were calculated using the Bio-rad CFX 

Manager software. 

Collection of field samples for screening using qPCR. Airborne sporangial 

concentrations were monitored during the cucumber growing season (May to 

September) in 2018 and 2019 using Burkard spore traps. Each year, a spore trap was 

placed 20 m from a commercial cucumber field located in Muskegon County in 

northwest Michigan and a cucumber research plot at the Michigan State University 

 

47 

 

 

(MSU) Plant Pathology Farm located in Ingham County in southcentral Michigan. The 

MSU cucumber research plot (0.25 ha) was direct seeded during the last week of July 

and was located 200 m from an abandoned hop research yard (0.25 ha) where 

systemically infected basal shoots (spikes) were observed beginning in late April 2019. 

An additional Burkard spore trap was placed in a commercial hop yard in Berrien 

County in 2019.  

The reel in each Burkard trap was covered with a melinex tape coated with an 

adhesive as described previously. The tape was removed weekly and cut longitudinally 

along the center line into two subsections of 9 x 336 mm each (Rogers et al., 2009). 

The first section was then cut into 48-mm lengths (equivalent to a monitoring period of 

24 h), scored at hourly intervals (2 mm) and stained to facilitate counting according to 

the protocol described by Granke et al. (2013). The second section was also cut into 48-

mm lengths, placed into impact-resistant 2mL tubes (Lysing Matrix H, MP Biomedicals) 

and subjected to DNA extraction as previously described using a NucleoSpin Plant II 

isolation kit (Macherey-Nagel, Bethlehem, PA). Subsequently, 2 µl of the extraction 

product was evaluated using qPCR. Fields were scouted weekly for signs and/or 

symptoms of P. cubensis. Leaf samples with lesions resembling CDM symptoms and 

signs of the pathogen were returned to the laboratory and examined using light 

microscopy to verify the presence of sporangia. .  

RESULTS 
 

Sensitivity and specificity of qPCR. Using 10-fold dilutions of P. cubensis and 

P. humuli DNA, the qPCR assay exhibited a significant  linear response with an 

efficiency of 93.6% (R2=0.99) and 90.7% (R2=1), respectively (Figs. 2-1A, 2-1C). Both 

 

48 

 

 

species-specific LNA probes detected each pathogen within total DNA template 

amounts ranging between 100 to 100,000 fg per reaction (Figs. 2-1A, 2-1C). The 

average Cq values for samples containing 100 fg of P. humuli and P. cubensis DNA was 

< 35.5. Most samples with concentrations below 100 fg were either not detected or 

detected without reasonable certainty (<95% of the times tested); thus, 100 fg of 

template DNA was considered as the lower limit of detection (LOD) of the qPCR assay 

for both species. Although the LNA probes were specifically designed to detect either P. 

humuli or P. cubensis based on the recognition of a SNP at the 105-base of the cox2 

gene (Summers et al., 2015), the HEX-labelled LNA probe CUBprobeSNP105 for P. 

cubensis detection showed nonspecific amplification of P. humuli DNA. However, the 

amplification curves of P. humuli DNA with this probe did not show the same shape as 

those generated using the P. cubensis DNA (Fig. 2-1B). P. cubensis samples were 

classified as positive based on the shape of the amplification curve using the probe 

CUBprobeSNP105 and no amplification of the FAM-labelled LNA probe 

HUMprobeSNP105. This probe (HUMprobeSNP105) designed to recognize P. humuli 

was highly specific and no background signal was observed when P. cubensis DNA 

was analyzed (Figs. 2-2D). 

Additionally, when mixed-samples containing DNA from both species were 

assessed using the qPCR assay, no significant changes were observed in the Cq 

values of the samples containing P. cubensis or P. humuli DNA in a 1:1 ratio (Table 2-

2). However, when 10,000 fg of P. cubensis DNA was mixed with 100 fg of P. humuli 

DNA, no detection of P. humuli DNA was observed (Table 2-2). Detection of P. 

cubensis occurred in all mixtures, but when 10,000 fg of P. humuli DNA was mixed with 

 

49 

 

 

1,000 or 100 fg of P. cubensis DNA, the detection of P. cubensis occurred at significant 

lower Cq values than in reactions including only P. cubensis DNA (Table 2-2). Mixed-

samples containing DNA from both species generated amplification curves with both 

probes (CUBprobeSNP105 and HUMprobeSNP105). This clearly differentiated them 

from samples containing only P. cubensis DNA for which there was only amplification 

with the CUBprobeSNP105 probe (Supplementary Fig. 2-S1 A, C). Although the 

samples containing only P. humuli DNA also generated amplification curves with both 

probes (Supplementary Fig. 2-S1 B), the amplification curves generated with the 

CUBprobeSNP105  probe for mixed samples showed faster growth in the exponential 

phase of the curves. This was the case for mixed-samples containing DNA from both 

species in a 1:1 ratio and samples containing P. cubensis and P. humuli DNA in 10:1 

ratio (Supplementary Fig. 2-S1 C, D). Only the amplification curves of the samples 

containing P. cubensis and P. humuli DNA in 1:10 and 1:100 ratio were not clearly 

differentiated from the amplification curves containing only P. humuli DNA 

(Supplementary Fig. 2-S1 B, E). 

Both LNA probes detected DNA from extractions containing 10, 100 or 1000 

sporangia of P. cubensis and P. humuli. Upon regression analysis, a linear relationship 

between Cq values and DNA extracted from purified sporangia was observed for both 

species (P. cubensis, R2= 0.99, Pvalue = 0.03 and P. humuli, R2= 0.98, Pvalue = 0.084) 

(Fig. 2-2A). The average Cq value for detecting 10 P. cubensis sporangia was < 35.5 

(Fig. 2-2A) and samples with less than 10 sporangia could not be detected with 

reasonable certainty (>95% of the times tested). Cq values ≤ 35.5 were classified as 

 

50 

 

 

specific to P. humuli and P. cubensis and were used as a threshold to evaluate field 

samples. 

The sensitivity of the assay to detect P. cubensis was minimally affected by the 

adhesive applied to the melinex tape (Fig. 2-2B). DNA was detected from extractions of 

tapes spiked with 20, 50,100, or 300 sporangia and a reduced number of samples 

(4/10) containing 10 sporangia had average Cq values below 35.5. All the extractions 

showed Cq values that were significantly different from the background signal observed 

in the negative controls. The relationship between sporangial numbers and Cq values 

was significant (p = 0.003) and the assay exhibited a linear response with a R2 value of 

0.99 (Fig. 2-2B). The average Cq values of the different sporangial dilutions were within 

the 95% confidence interval. However, a high standard error of the mean was observed 

among biological replicates of extractions with the same number of sporangia (Fig. 2-

2B) indicating that the extraction affects the precision of quantifying sporangia using the 

qPCR-based assay. 

Assessment of field samples using light microscopy and qPCR. A total of 

560 samples collected from May to August in 2018 and 2019 were assessed using 

qPCR. P. cubensis or P. humuli DNA was detected using qPCR on field samples with 

fewer than 10 sporangia (Fig. 2-2C). Approximately 90% of all samples that tested 

positive for either pathogen using qPCR (204 out of 227) had one or more sporangia on 

the corresponding half of the tape analyzed using light microscopy. The average Cq 

value of the IC remained relatively constant and had an average of 28.8 ± 1.7 (SD) 

among all the field samples evaluated. Regression analysis indicated that the number of 

sporangia on the second half of the tape of the samples (quantified using light 

 

51 

 

 

microscopy) explained 54% (R2 = 0.54) and 10% (R2 = 0.10) of the variation in the Cq 

values of P. cubensis and P. humuli, respectively (Fig. 2-2C).  

Using light microscopy, Pseudoperonospora spp. sporangia were first detected in 

2018 during May (Muskegon Co. commercial field, 15 May) and June (Ingham Co. 

research field, June 13) (Fig. 2-3). From May to July at the research field, fewer than 5 

sporangia/day were observed via light microscopy with P. cubensis DNA confirmed 

using qPCR on 13 June and 10, 23 July; while P. humuli was confirmed on 19 June and 

7 July (Fig. 2-3A). At the commercial field, fewer than 10 sporangia/day were detected 

using light microscopy from May to July, except for 21 June (Fig. 2-3B). Using qPCR, 

we confirmed the presence of P. humuli DNA on 29 May to 4 June and 21 to 26 June 

(Fig. 2-3B) while P. cubensis DNA was detected on 5 and 13 June (Fig. 2-3B). Airborne 

sporangial concentrations increased during the first week of August and reached the 

maximum during the third week of the month in both locations monitored in 2018 (Figs. 

2-3A, 2-3B). CDM symptoms were observed at the research field on 15 August, a peak 

of 16 sporangia was observed via light microscopy 10 days prior (5 August) and P. 

cubensis DNA was confirmed by qPCR on 1 to 6 August and 10 August (Fig. 2-3A). 

After CDM symptoms were observed in the research field, daily sporangial counts via 

light microscopy exceeded 10 sporangia/day with P. cubensis DNA detected nearly 

every day using qPCR (Figs. 2-3A, 2-3B). CDM symptoms were observed at the 

commercial field on 7 August, a peak of 22 sporangia was observed via light 

microscopy four days prior (3 August) and P. cubensis DNA was confirmed by qPCR on 

3, 5 August, and 28, 30, and 31 July (Fig. 2-3B). The day following the detection of 

 

52 

 

 

CDM symptoms, more than 80 sporangia/day were captured by the spore traps and P. 

cubensis DNA was detected every day using qPCR (Fig. 2-3B). 

In 2019, Pseudoperonospora spp. sporangia were first detected in May across all 

locations using light microscopy (Fig. 2-4). During May and June, concentrations of 

airborne sporangia exceeded 10 sporangia/day in the commercial hop yard (Fig. 2-4A) 

and the research field (Fig. 2-4B). Using qPCR, P. humuli DNA was detected several 

times in the commercial hop yard from May through August while P. cubensis DNA was 

only detected on 12, 14, and 18 August (Fig. 2-4A). Based on data from light 

microscopy, the research field which was in proximity to a non-treated hop yard, had 

more than 10 sporangia/day during the last week of May, the first and fourth week of 

June and the first week of July (Fig. 2-4B). At this location, P. humuli DNA was detected 

from May to July using qPCR (Fig. 2-4B). During August, fewer than 10 sporangia/day 

were observed at the research field. CDM symptoms were confirmed at this site on 21 

August and P. cubensis DNA was verified by qPCR on 11, 19, and 20 August. Following 

CDM symptoms, P. cubensis DNA was detected from 22 to 31 August (Fig. 2-4B). In 

the commercial cucumber field, fewer than 10 sporangia/day were observed using light 

microscopy from May through July with the exception of 18, 21 and 23 June (Fig. 2-4C). 

P. humuli DNA was confirmed with qPCR on 22 and 23 June (Fig. 2-4C). At this 

location, the sporangial counts increased from the third to the last week of August. CDM 

symptoms were confirmed on 16 August and P. cubensis DNA was detected using 

qPCR in air samples every day from 12 to 31 August (Fig. 2-4C). 

 

53 

 

 

 

DISCUSSION 
 

The ability to detect and differentiate between P. cubensis and P. humuli in field 

air samples using qPCR represents an important advance for CDM monitoring and 

management. The qPCR detection of airborne sporangia could be used as a decision-

making tool to initiate fungicide sprays (Dhar et al., 2019) or as a complementary 

variable to forecast the risk  (Carisse et al., 2009) of CDM outbreaks in Michigan. Early 

and specific detection of P. cubensis sporangia could ensure timely crop protection and 

avoid unnecessary fungicide applications. P. cubensis does not overwinter in Michigan 

and for disease to occur, the pathogen must be introduced into the state’s growing 

regions annually (Naegele et al., 2016). Burkard spore traps coupled with light 

microscopy have been used since 2007 to alert Michigan growers to an influx of P. 

cubensis sporangia into their growing region. However, HDM is prevalent in the state 

(Lizotte et al., 2020) where approximately 400 ha of hops have been planted (Hop 

Growers of America, 2019). Using a qPCR assay, we were able to distinguish between 

the morphologically similar sporangia of P. cubensis and P. humuli collected from 

Burkard spore traps. During the two years of monitoring using Burkard spore traps 

coupled with the qPCR assay in cucumber fields, P. cubensis sporangia were detected 

approximately 5-10 days before CDM symptoms were observed in monitored cucumber 

fields.  

We adapted the qPCR assay developed by Summers et al. (2015) to a high 

degree of sensitivity for use with the Burkard spore trap samples. Using DNA extracted 

from purified sporangial suspensions of P. cubensis and P. humuli, we were able to 

 

54 

 

 

detect DNA concentrations ranging from 100 fg to 100,000 fg.  This sensitivity was 

validated with the detection of the two downy mildew pathogens in field samples 

containing less than 10 sporangia (Cq <35.5). We split the tape of Burkard spore traps 

to facilitate the comparison between light microscopy and qPCR and observed that the 

number of sporangia deposited onto one half of the tape was linearly correlated with the 

Cq values obtained after the assessment of the other half using qPCR. However, the 

change in the number of sporangia on field samples quantified using light microscopy 

explained only 54% and 10% of the variation in the Cq values of P. cubensis and P. 

humuli, respectively, suggesting a limited capacity of the qPCR assay for the absolute 

quantification of sporangia in field samples.  

The low correlation between Cq values and sporangial numbers of field samples 

may be explained by the high variation in the yield of DNA extraction among samples 

(Summers et al., 2015), inaccurate visual quantification, the low specificity of one of our 

probes, and possibly, the multicopy nature of the target sequence (Klosterman et al., 

2014; Kunjeti et al., 2016; Dung et al., 2018). Similarly, the yield variation of DNA 

extractions among samples may also explain the variation observed in the Cq values of 

samples with the same number of sporangia in vitro. This variation is introduced in all 

the samples collected in the field and may reduce the precision for the quantification of 

sporangia using the extraction protocol and qPCR assay described in this study. 

However, assessing the first half of the spore trap tape using qPCR could reduce the 

number of samples that require microscopic analysis for spore quantification, 

accelerating the turn-around time associated with monitoring airborne P. cubensis 

sporangia. The reduction of variation in the yield of DNA extractions and the utilization 

 

55 

 

 

of a qPCR assay based on a single-copy marker may be more appropriate for 

quantification (Rahman et al., 2020) but may result in a system with reduced sensitivity 

when compared to the multi-copy system that we used.  

The nonspecific amplification of P. humuli DNA affected the quantification 

capacity of the assay when both species were present in the same reaction. However, 

the inclusion of a second probe ensured that the detection of each species was possible  

even when P. cubensis and P. humuli were present in the same sample. The 

nonspecific amplification of P. humuli did not occur under the conditions described by 

Summers et al. (2015) and was a consequence of the change in the commercial master 

mix used for the qPCR reactions (Supplementary Fig. 2-S2). The master mix used in 

this study reduced the variation among technical replicates (data not shown) and 

increased the amplification efficiency of the qPCR reactions (exponential phase) but 

affected the specificity of the assay. Different qPCR master mixes influences how 

oligonucleotides (primers and probes) bind to target regions (Morinha et al., 2020), thus 

the suitability of new reagents must be carefully evaluated as they may condition the 

results of the qPCR. The detection of both species using our qPCR assay was hindered 

only in samples with a significantly higher amount of P. humuli compared to P. cubensis 

(i.e. samples containing P. cubensis and P. humuli DNA in 1:10 and 1:100 ratio). In 

these cases, or in locations where a higher number of P. humuli sporangia relative to 

the number of P. cubensis sporangia is expected (e.g. hop yards) the use of the IQ 

Supermix (Bio-rad, Hercules, CA) as described by Summers et al. (2015) for qPCR 

reactions should allow a more accurate evaluation of the samples. 

 

56 

 

 

During two years of monitoring in commercial cucumber fields, we did not detect 

any periods when both pathogens were detected simultaneously, however, overlapping 

periods may have occurred at the commercial hop yard during late August. The 

identification of genetic regions with a higher number of polymorphisms has allowed the 

design of more specific primer and probes for P. cubensis detection (Rahman et al., 

2020). Using this new set of primers and probes in combination with the probe 

HUMprobeSNP105 (for detection of P. humuli) could ensure both specific detection and 

quantification of P. cubensis and P. humuli sporangia using qPCR even during periods 

when both species are present. 

Despite the limitations of the qPCR assay described in this study, we were able 

to detect low atmospheric concentrations of P. cubensis and P. humuli (<10 

sporangia/day). Detection of P. cubensis before symptoms developed in the field was 

linked to a sporangial concentration below 10 sporangia/day as estimated using light 

microscopy. In other crops including lettuce and onion, measurements of aerial spore 

load (sporangia/day) have been used to guide fungicide application to control Bremia 

lactucae (Dhar et al., 2019) and Botrytis squamosa (Carisse et al., 2009), respectively. 

In these systems, fungicide applications began once spore loads reached a critical level 

between 300-500 spores/day (10 sporangia/m3). In the cucumber fields monitored in 

2018, CDM symptoms were observed after airborne P. cubensis sporangial 

concentrations were greater than 15 sporangia/day as determined via light microscopy, 

suggesting that the critical concentration for CDM could be close to this number 

depending on the coincident environmental conditions. Using the qPCR assay, P. 

cubensis sporangia were detected before concentrations reached >15 sporangia/day. 

 

57 

 

 

More than 15 sporangia/day were also observed one month before CDM was detected 

in the cucumber fields (June 2018 and 2019), however, these sporangia were identified 

as P. humuli using qPCR. In Michigan, information on the airborne concentration of 

sporangia is used to provide an early warning for growers that CDM is likely to occur 

(Hausbeck, 2020) and prompt the application of fungicides. Using light microscopy only, 

P. humuli sporangia could have triggered unnecessary fungicide applications, 

highlighting the importance of a qPCR assay system that reliably distinguishes between 

P. cubensis and P. humuli. 

Using light microscopy and qPCR, differences in the airborne sporangial 

concentrations of P. humuli and P. cubensis were detected between the two years of 

monitoring. From June to August 2018 at the commercial cucumber field, we detected 

higher airborne sporangial concentrations of P. cubensis compared to 2019. A relatively 

cold and rainy spring delayed the planting of cucumbers for pickling in 2019  (USDA, 

2020). This reduced host availability may have also resulted in reduced infection and P. 

cubensis sporangia production. Similarly, from May to July 2018, low concentrations of 

P. humuli (<10 sporangia/day) were detected in the two monitored fields whereas a 

higher concentration (>10 sporangia/day) was observed at the three locations monitored 

in 2019. P. humuli overwinters in dormant hop buds or crowns, growing into expanding 

basal shoots in spring and early summer (Coley-Smith, 1962). Extended periods of 

wetness, high RH, and temperatures below 20°C (Royle, 1973; Gent and Ocamb, 2009) 

occurred during the cold and rainy spring of 2019 (NOAA, 2019) and may have favored 

the pathogen’s reproduction and infection. 

 

58 

 

 

These results suggest that the qPCR-based assay allowed for precise monitoring 

of airborne P. cubensis and P. humuli sporangia over two different years; the specific 

detection of these two species was not possible using light microscopy only. Cucumber 

growers in Michigan desire to know when sporangia of P. cubensis have arrived in their 

production region/field so that scouting efforts can be intensified, and costly fungicide 

programs initiated. The information on P. cubensis detection derived from spore traps 

coupled with qPCR could be used by growers to make informed decisions regarding 

fungicide usage leading to increased efficiency. Judicious use of fungicides may slow 

the development of pathogen resistance and decrease the cost associated with CDM 

control. The deployment of a broader network of spore traps and the evaluation of air 

samples using qPCR could also improve the risk assessment of CDM epidemics. Future 

evaluation of more cost-effective spore traps such as impaction traps for the monitoring 

of P. cubensis is essential to increase the geographic coverage of the spore trapping 

network in Michigan. The use of more spore traps at a local level could make the 

monitoring more geographically precise and trigger the execution of disease 

management practices only in fields at high risk of infection based on the qPCR 

detection of P. cubensis and the local environmental conditions. 

 

 

59 

 

 

 
 
 
 
 
 
 
 
 
 
 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

APPENDIX 

60 

 

 

 

 
 
Table 2-1. Primers and locked nucleic acids (LNA) probes for the qPCR assay differentiating Pseudoperonospora 
cubensis and P. humuli using the 105 SNP in the mitochondrial Cox2 gene. 

APPENDIX 

 

Code name  

[Conc]a  

Sequence 5'->3' 

Primer  RT33Fb 

Primer  RT182Rb 

600 nM  

600 nM  

AACTCCCGTTATGGAAGGTATT 

CCATGTACAACAGTAGCTGGA 

Probe  CUBprobeSNP105b  250 nM 

HEX/A+C+AAA+C+G+AATA+CT/BHQ c 

Probe  HUMprobeSNP105b  500 nM 

FAM/AA+C+AAA+C+A+AATA+CTG/BHQ c 

Probe 

ICprobeJ2 

250 nM 

CYS/A+GCATTATT+GTTTAT+CATATATACA/BHQc 

Internal 

Control  

IC 

7.5 x10-10 nM 

GATTTGTATATATGATAAACAATAATGCTATAAC 

(0.75 aMd) 

ATAGAGTCTCTTTCATGAATAATCCAGCTACTGT 

AACTCCCGTTATGGAAGGTATTATCATTAATCAT 

TGTACATGG 

aConcentrations used in a 20 µl qPCR reaction.  
bPrimers and LNA probes were adapted from Summers et al. (2015).  
cLocked nucleic acids in the probes are followed by a plus (+) sign 
dAttomole (aM) = 10-18 moles per liter 

 

 

61 

 

Table 2-2. Threshold cycle (Cq) values of the qPCR assays using LNA probes and 
varying concentrations of genomic DNA. 

DNA 

Probe 

CUBprobeSNP105  HUMprobeSNP105 

Cq-HEX 

SD 

Cq-FAM 

  

P. cubensis   P. humuli 

10,000 fg 

1,000 fg 

Un-

100 fg 

mixed 

-- 

-- 

-- 

-- 

-- 

-- 

28.39a 

32.12b 

35.49c 

10,000 fg 

29.88nc 

1,000 fg 

32.96nc 

100 fg 

NA 

10,000 fg 

10,000 fg 

28.67a 

1,000 fg 

1,000 fg 

31.44b 

100 fg 

100 fg 

35.32c 

Mixed 

10,000 fg 

1,000 fg 

28.37a 

10,000 fg 

100 fg 

28.39a 

1,000 fg 

10,000 fg 

29.78x 

100 fg 

10,000 fg 

30.58y 

0.10 

0.21 

0.74 

0.34 

0.17 

NA 

0.23 

0.37 

0.98 

0.15 

0.26 

0.37 

0.24 

NA 

NA 

NA 

28.29d 

31.70e 

36.61f 

28.52d 

32.42e 

35.11f 

32.07e 

NAz 

27.96d 

24.99d 

SD 

NA 

NA 

NA 

0.08 

0.30 

0.14 

0.60 

0.08 

0.59 

0.33 

NA 

0.34 

5.91 

The HEX-labelled probe (CUBprobeSNP105) was designed to detect only DNA from P. 
cubensis and the FAM-labelled probe (HUMprobeSNP105) was designed to detect only 
DNA from P. humuli. Cq values with the same letter are not significantly different (t-Test; 
P=0.05). --: not DNA added. NC: not used for comparison in the t-Test. NA: not defined.  
 
 
 
 
 
 

 

62 

 

 

 
 

 
Figure 2-1. Regression and amplification curves of Pseudoperonospora cubensis 
and P. humuli DNA using qPCR.  
A. Standard curve for the quantification of P. cubensis and P. humuli DNA using the 
LNA probe CUBprobeSNP105. The log10 of DNA (100fg, 1,000fg, 10,000fg, and 
100,000fg) is plotted against the quantification cycle (Cq) values. Each curve was 
plotted separately using DNA from each pathogen. The data points below 100fg were 
not included in the regression analysis.  B. Amplification curves of P. cubensis and P. 
humuli DNA with different concentrations using the LNA probe CUBprobeSNP105. Each 
curve was plotted separately using DNA from each pathogen C. Standard curve for the 
quantification of P. cubensis and P. humuli DNA using the LNA probe 
HUMprobeSNP105. The log10 of DNA (100fg, 1,000fg, 10,000fg, and 100,000fg) is 
plotted against the quantification cycle (Cq) values. Each curve was plotted separately 
using DNA from each pathogen. D. Amplification curves of P. cubensis and P. humuli 
DNA with different concentrations using the LNA probe HUMprobeSNP105. Each curve 
was plotted separately using DNA from each pathogen 

 

63 

 

 

Figure 2-2. Standard curves for the quantification of Pseudoperonospora cubensis and P. humuli sporangia 
using qPCR.  
A. The log10 of the number of sporangia is plotted against the quantification cycle values (Cq). The centerline represents 
the line of fit and error bars represent standard error of the mean. Each curve was plotted separately using the LNA 
probes specific to each pathogen. Data points represent three technical replicates from two DNA extractions. B. Standard 
curves based on the qPCR assays of DNA extractions from P. cubensis sporangia (20, 50, 100 and 300) in the presence 
and absence of the adhesive mix used on the melinex tape. All data points are from three technical replicates from 4 
independent DNA extractions. C.  Linear regression of Pseudoperonospora spp. sporangia counted using light 
microscopy against corresponding mean Cq values. All the samples used in this regression were collected using spore 
traps in the field.

 

 

 

64 

 

 
Figure 2-3. Monitoring of Pseudoperonospora cubensis and P. humuli sporangia 
using Burkard spore traps in Ingham (A) and Muskegon (B) in 2018. 
The data from each county was divided into two panels. The first top panel represents 
the daily sporangia numbers estimated through the analysis of Burkard spore trap 
samples using light microscopy (blue bars). The y-axis was trimmed to 40 sporangia to 
facilitate the visualization of low counts. The middle panel represents the qPCR results 
for the detection of P. cubensis (red bars) and P. humuli (green bars) in the tape of 
Burkard spore traps. Yellow triangles denote the monitoring starting date. Red triangles 
denote the first confirmed report of cucurbit downy mildew in the state. The dashed line 
denotes the date of cucurbit downy mildew detection in the field. Scouting efforts to 
detect CDM symptoms in growing cucumber regions are intensified once sporangial 
loads exceed 10 sporangia/day.   
 
 

 

65 

 

 

 
Figure 2-4. Monitoring of Pseudoperonospora cubensis and P. humuli sporangia 
using Burkard spore traps in Berrien (A), Ingham (B) and Muskegon (C) in 2019.  
The data from each county was divided into two panels. The first top panel represents 
the daily sporangia numbers estimated through the analysis of Burkard spore trap 
samples using light microscopy (blue bars). The y-axis was trimmed to 40 sporangia to  
 

 

66 

 

 

Figure 2-4. (cont’d) 
 
facilitate the visualization of low counts. The middle panel represents the qPCR results 
for the detection of P. cubensis (red bars) and P. humuli (green bars) in the tape of 
Burkard spore traps. Yellow triangles denote the monitoring starting date. Red triangles 
denote the first confirmed report of cucurbit downy mildew in the state. The dashed line 
denotes the date of cucurbit downy mildew detection in the field. Scouting efforts to 
detect CDM symptoms in growing cucumber regions are intensified once sporangial 
loads exceed 10 sporangia/day.   
 

 

67 

 

 

A

M
A
F
U
F
R

 

20000

10000

0

30
Cycles
B

20000

M
A
F
U
F
R

 

10000

0

30
Cycles
C

20000

M
A
F
U
F
R

 

10000

0
35

30
Cycles
D

20000

10000

M
A
F
U
F
R

 

0

35

30
Cycles
E

20000

10000

M
A
F
U
F
R

 

0

35

30
Cycles

A

M
A
F
U
F
R

 

20000

10000

0

B

20000

10000

M
A
F
U
F
R

 

0

C

M
A
F
U
F
R

 

D

M
A
F
U
F
R

 

E

M
A
F
U
F
R

 

20000

10000

0

20000

10000

0

20000

10000

0

25

25

25

25

25

Sample

Phu_0fg__Pcu_10000fg
Phu_0fg__Pcu_1000fg
Phu_0fg__Pcu_100fg

X
E
H
U
F
R

 

Phu_0fg__Pcu_10000fg
Phu_0fg__Pcu_1000fg
Phu_0fg__Pcu_100fg

1000

750
Sample
500

250

35

25

40

30
Cycles

35

40

0

1000

25

750
Sample

Sample

Phu_10000fg__Pcu_0fg
Phu_1000fg__Pcu_0fg
Phu_100fg__Pcu_0fg

X
E
H
U
F
R

 

500

Phu_10000fg__Pcu_0fg
Phu_1000fg__Pcu_0fg
Phu_100fg__Pcu_0fg

250

35

25

40

30
Cycles

35

40

0

1000

25

Sample

Phu_10000fg__Pcu_10000fg
Phu_1000fg__Pcu_1000fg
Phu_100fg__Pcu_100fg

750

Sample

X
E
H
U
F
R

 

500

Phu_10000fg__Pcu_10000fg
Phu_1000fg__Pcu_1000fg
Phu_100fg__Pcu_100fg

250

25

40

30
Cycles

35

40

0

1000

25

Sample

Phu_1000fg__Pcu_10000fg
Phu_100fg__Pcu_10000fg

750
Sample
500
Phu_1000fg__Pcu_10000fg
Phu_100fg__Pcu_10000fg
250

X
E
H
U
F
R

 

25

40

30
Cycles

35

40

0

1000

25

Sample

Phu_10000fg__Pcu_1000fg
Phu_10000fg__Pcu_100fg

750
Sample
500
Phu_10000fg__Pcu_1000fg
Phu_10000fg__Pcu_100fg
250

X
E
H
U
F
R

 

0

1000

750

500

250

0

X
E
H
U
F
R

 

30
Cycles
1000

X
E
H
U
F
R

 

750

500

250

0

30
Cycles
1000

X
E
H
U
F
R

 

750

500

250

0

30
Cycles
1000

X
E
H
U
F
R

 

750

500

250

0

30
Cycles
1000

X
E
H
U
F
R

 

750

500

250

0

Sample

Sample

P. cub

Sample

P. hum
0 fg
0 fg
0 fg

10,000 fg
Phu_0fg__Pcu_10000fg
Phu_0fg__Pcu_1000fg
1,000 fg
Phu_0fg__Pcu_100fg
100fg

Phu_0fg__Pcu_10000fg
Phu_0fg__Pcu_1000fg
Phu_0fg__Pcu_100fg

35

25

40

30
Cycles

35

40

Sample

Sample

Phu_10000fg__Pcu_0fg
Phu_1000fg__Pcu_0fg
Phu_100fg__Pcu_0fg

Sample

P. hum
10,000 fg
1,000 fg
100 fg

Phu_10000fg__Pcu_0fg
Phu_1000fg__Pcu_0fg
Phu_100fg__Pcu_0fg

P. cub
0 fg
0 fg
0fg

35

25

40

30
Cycles

35

40

Sample

Sample

Phu_10000fg__Pcu_10000fg
Phu_1000fg__Pcu_1000fg
Phu_100fg__Pcu_100fg

P. cub

P. hum
Sample
10,000 fg
1,000 fg
100 fg

10,000 fg
Phu_10000fg__Pcu_10000fg
Phu_1000fg__Pcu_1000fg
1,000 fg
Phu_100fg__Pcu_100fg
100fg

35

25

40

30
Cycles

35

40

Sample

Sample

P. cub

P. hum
Sample
1,000 fg
100 fg

10,000 fg
Phu_1000fg__Pcu_10000fg
1,0000 fg
Phu_100fg__Pcu_10000fg

Phu_1000fg__Pcu_10000fg
Phu_100fg__Pcu_10000fg

35

25

40

30
Cycles

35

40

Sample

Sample

P. hum
Sample
10,000 fg
10,000 fg

P. cub
1,000 fg
Phu_10000fg__Pcu_1000fg
100 fg
Phu_10000fg__Pcu_100fg

Phu_10000fg__Pcu_1000fg
Phu_10000fg__Pcu_100fg

25

40

30
Cycles

35

40

25

30
Cycles

35

25

40

30
Cycles

35

40

Supplementary Figure S1. Amplification curves of Pseudoperonospora cubensis (Pcu) and P. humuli 
Supplementary Figure 2-S1. Amplification curves of Pseudoperonospora cubensis 
(Phu) DNA using qPCR. A. Amplification curves of P. cubensis DNA with different concentrations using 
the LNA probes HUMprobeSNP105 (left, FAM) and CUBprobeSNP105 (right. HEX). B. Amplification 
(Pcu) and P. humuli (Phu) DNA using qPCR. A. Amplification curves of P. cubensis 
curves of P. humuli DNA with different concentrations using the LNA probes HUMprobeSNP105 (left, 
DNA with different concentrations using the LNA probes HUMprobeSNP105 (left, FAM) 
FAM) and CUBprobeSNP105 (right, HEX). C. Amplification curves of P. humuli and P. cubensis DNA mix 
and CUBprobeSNP105 (right. HEX). B. Amplification curves of P. humuli DNA with 
in a 1:1 ratio using the LNA probes HUMprobeSNP105 (left, FAM) and CUBprobeSNP105 (right, HEX). 
different concentrations using the LNA probes HUMprobeSNP105 (left, FAM) and 
D. Amplification curves of P. humuli and P. cubensis DNA mix in a 1:10 and 1:100 ratio using the LNA 
probes HUMprobeSNP105 (left, FAM) and CUBprobeSNP105 (right, HEX). E. Amplification curves of P. 
CUBprobeSNP105 (right, HEX).  
cubensis and P. humuli DNA mix in a 1:10 and 1:100 ratio using the LNA probes HUMprobeSNP105 (left, 
 
FAM) and CUBprobeSNP105 (right, HEX) 

 

 

 

68 

 

 

Supplementary Figure 2-S1. (cont’d) 
 
C. Amplification curves of P. humuli and P. cubensis DNA mix in a 1:1 ratio using the 
LNA probes HUMprobeSNP105 (left, FAM) and CUBprobeSNP105 (right, HEX). D. 
Amplification curves of P. humuli and P. cubensis DNA mix in a 1:10 and 1:100 ratio 
using the LNA probes HUMprobeSNP105 (left, FAM) and CUBprobeSNP105 (right, 
HEX). E. Amplification curves of P. cubensis and P. humuli DNA mix in a 1:10 and 
1:100 ratio using the LNA probes HUMprobeSNP105 (left, FAM) and 
CUBprobeSNP105 (right, HEX)  

 

 

 

 

 

 

 

 

 

69 

 

 

A

25000

M
A
F
U
F
R

 

20000

15000

10000

5000

0

C

3000

M
A
F
U
F
R

 

2000

1000

0

20

25

30

Cycles

35

40

Species

P_cubensis
P_humuli

fg

10000fg
1000fg
100fg

Species

P_cubensis
P_humuli

fg

10000fg
1000fg
100fg

B

600

X
E
H
U
F
R

 

400

200

0

D

200

X
E
H
U
F
R

 

100

0

20

25

30

Cycles

35

40

Species

P_cubensis
P_humuli

fg

10000fg
1000fg
100fg

Species

P_cubensis
P_humuli

fg

10000fg
1000fg
100fg

20

25

35

40

30

Cycles

20

25

35

40

30

Cycles

 
 
Supplementary Figure S2. Amplification curves of Pseudoperonospora cubensis and P. humuli DNA 
using qPCR. A. Amplification curves of P. cubensis and P. humuli DNA with different concentrations using 
the LNA probe HUMprobeSNP105 and the Prime-Time Gene Expression Master Mix (IDT, Skokie, IL). B. 
Supplementary Figure 2-S2. Amplification curves of Pseudoperonospora cubensis and 
Amplification curves of P. cubensis and P. humuli DNA with different concentrations using the LNA probe 
P. humuli DNA using qPCR. A. Amplification curves of P. cubensis and P. humuli DNA 
CUBprobeSNP105 and the Prime-Time Gene Expression Master Mix. C. Amplification curves of P. 
with different concentrations using the LNA probe HUMprobeSNP105 and the Prime-
cubensis and P. humuli DNA with different concentrations using the LNA probe HUMprobeSNP105 and 
Time Gene Expression Master Mix (IDT, Skokie, IL). B. Amplification curves of P. 
the IQ Supermix (Bio-rad, Hercules, CA). D. Amplification curves of P. cubensis and P. humuli DNA with 
different concentrations using the LNA probe CUBprobeSNP105 and the IQ Supermix. 
cubensis and P. humuli DNA with different concentrations using the LNA probe 
 
CUBprobeSNP105 and the Prime-Time Gene Expression Master Mix. C. Amplification 
curves of P. cubensis and P. humuli DNA with different concentrations using the LNA 
probe HUMprobeSNP105 and the IQ Supermix (Bio-rad, Hercules, CA). D. Amplification 
curves of P. cubensis and P. humuli DNA with different concentrations using the LNA 
probe CUBprobeSNP105 and the IQ Supermix.  

 

 

 

 

70 

 

 

 

 

 

 

 

 

 

 

 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

LITERATURE CITED 

71 

 

 

LITERATURE CITED 

 
 
Adams, M. L., and Quesada-Ocampo, L. M. (2014) Evaluation of fungicides for control 
of downy mildew on cucumber. Kinston 2013. Plant Dis Manag Rep 8: V240 
 
Blum, M., Waldner, M., Olaya, G., Cohen, Y., Gisi, U., and Sierotzki, H. (2011) 
Resistance mechanism to carboxylic acid amide fungicides in the cucurbit downy 
mildew pathogen Pseudoperonospora cubensis. Pest Manag Sci 67: 1211–1214 
 
Brzozowski, L., Holdsworth, W., and Mazourek, M. (2016) ‘DMR-NY401’: A new downy 
mildew–resistant slicing cucumber. HortScience 51: 1294–1296 
 
Bustin, S. A., Benes, V., Garson, J.A., Hellemans, J., Huggett, J., Kubista, M., Mueller, 
R., Nolan, T., Pfaffl, M.W., Shipley, G.L., Vandesompele, J., and Wittwer, C. T (2009) 
The MIQE guidelines: minimum information for publication of quantitative real-time PCR 
experiments. Clin Chem 55: 611–622 
 
Carisse, O., Tremblay, D.M., Lévesque, C.A., Gindro, K., Ward, P., and Houde, A. 
(2009) Development of a Taqman real-time PCR assay for quantification of airborne 
conidia of Botrytis squamosa and management of Botrytis leaf blight of onion. 
Phytopathology 99: 1273–1280 
 
Cohen, Y., Van den Langenberg, K.M., Wehner, T.C, Ojiambo, P.S., Hausbeck, M, 
Quesada-Ocampo, L.M., Lebeda, A., Sierotzki, H., and Gisi, U. (2015) Resurgence 
of Pseudoperonospora cubensis: the causal agent of cucurbit downy mildew. 
Phytopathology 105: 998–1012 
 
Cohen, Y., and Rotem, J. (1971) Dispersal and viability of sporangia 
of Pseudoperonospora cubensis. Trans Br Mycol Soc 57: 67–74 
 
Coley-Smith, J.R. (1962) Overwintering of hop downy mildew Pseudoperonospora 
humuli (Miy. and Tak.) Wilson. Ann Appl Biol 50: 235–243 
 
Dhar, N., Mamo, B.E., Subbarao, K., Koike, S., Fox, A., Anchieta, A., and Klosterman, 
S.J. (2019) Measurements of aerial spore load by qPCR facilitates lettuce downy 
mildew risk advisement. Plant Dis 104: 82–93 
 
Dung, J.K.S, Scott, J.C., Cheng, Q., Alderman, S.C., Kaur, N., Walenta, D.L., Frost, 
K.E., and Hamm, P.B. (2018) Detection and quantification of airborne Claviceps 
purpurea sensu lato ascospores from Hirst-type spore traps using real-time quantitative 
PCR. Plant Dis 102: 2487–2493 
 
Ernest, E.G., Kee, W.E., and Everts, K.L. (2005) Evaluation of fungicides for control of 
downy mildew on pickling cucumber, Trial 2, 2005. Plant Dis Manag Rep 2: V066 
 

 

72 

 

 

Gent, D.H., Ocamb, C.M. (2009) Predicting infection risk of hop by Pseudoperonospora 
humuli. PhytopathologyTM 99: 1190–1198 
 
Gent, D.H., Nelson, M.E., Farnsworth, J.L. and Grove, G.G. (2009), PCR detection of 
Pseudoperonospora humuli in air samples from hop yards. Plant Pathology, 58: 1081-
1091. 
 
Gevens, A.J., and Hausbeck, M.K. (2005) Control of downy mildew of cucumbers with 
fungicides, 2005. Plant Dis Manag Rep 2: V066 
 
Granke, L.L., and Hausbeck, M.K. (2011) Dynamics of Pseudoperonospora 
cubensis sporangia in commercial cucurbit fields in Michigan. Plant Dis 95: 1392–1400 
 
Granke, L.L., Morrice, J.J., and Hausbeck, M.K. (2013) Relationships between 
airborne Pseudoperonospora cubensis sporangia, environmental conditions, and 
cucumber downy mildew severity. Plant Dis 98: 674–681 
 
Gugino, B.K., and Grove, T.L. (2016) Evaluation of selected fungicides and fungicide 
programs for managing downy mildew of cucumber, 2015. Plant Dis. Manag. Rep. 10: 
V072: 
 
Hausbeck, M. K., and Linderman, S.D. (2014) Evaluation of fungicides for control of 
downy mildew of cucumber, 2013. Plant Dis Manag Rep 8: V304 
 
Hausbeck, M., Brisco-McCann, E., Goldenhar, K., and Cook, A. (2017) Evaluation of 
fungicide programs for control of downy mildew of cucumber, 2016. Plant Dis Manag 
Reports 11: V063 
 
Hausbeck, M.K., and Cortright, B. (2010) Evaluation of fungicides and application 
methods for control of downy mildew of cucumber, 2010. Plant Dis. Manag. Rep. V6: 
 
Hausbeck, M.K., Perla, D.E., and Cook, A.J. (2019) Evaluation of single fungicide 
products for control of downy mildew of cucumber, 2018. Plant Dis Manag Reports 139: 
V13 
 
Hausbeck, M.K. (2020) First cucumber downy mildew outbreak of 2020 confirmed in 
Berrien County. MSU Ext., https://www.canr.msu.edu/news/cucumber-downy-mildew-
outbreak-of-2020 
 
Hirst, J.M. (1952) An automatic volumetric spore trap. Annals of Applied Biology, 39, 
259–265. 
 
Holmes, G.J., Ojiambo, P.S., Hausbeck, M.K., Quesada-Ocampo, L., and Keinath, A.P. 
(2014) Resurgence of cucurbit downy mildew in the united states: a watershed event for 
research and extension. Plant Dis 99: 428–441 
 

 

73 

 

 

Hop growers of America (2019) 2018 Statistical report - hops. HGA Stat. Pack. - Revis., 
https://www.usahops.org/img/blog_pdf/168.pdf 
 
Kanetis, L., Holmes, G.J, and Ojiambo, P.S. (2010) Survival of Pseudoperonospora 
cubensis sporangia exposed to solar radiation. Plant Pathol 59: 313–323 
 
Keinath, A. P (2015) Utility of a cucumber plant bioassay to assess fungicide efficacy 
against Pseudoperonospora cubensis. Plant Dis 100: 490–499 
 
Klosterman, S.J., Anchieta, A., McRoberts, N., Koike, S.T., Subbarao, K. V., Voglmayr, 
H., Choi, Y-J., Thines, M., and Martin, F.N. (2014) Coupling spore traps and quantitative 
PCR assays for detection of the downy mildew pathogens of spinach (Peronospora 
effusa) and beet (P. schachtii). Phytopathology 104: 1349–1359 
 
Kunjeti, S.G., Anchieta, A., Martin, F.N., Choi, Y-J., Thines, M., Michelmore, R.W.,  
Koike, S.T., Tsuchida, C., Mahaffee, W., Subbarao, K. V., Klosterman, S. J. (2016) 
Detection and quantification of Bremia lactucae by spore trapping and quantitative PCR. 
Phytopathology 106: 1426–1437 
 
Lebeda, A., Cohen, Y. (2011) Cucurbit downy mildew (Pseudoperonospora cubensis)—
biology, ecology, epidemiology, host-pathogen interaction and control. Eur J Plant 
Pathol 129, 157–192 
 
Langston, D., Sanders, F.H. (2013) Evaluation of fungicides for control of downy mildew 
on cucumber in Georgia I, 2012. Plant Dis. Manag. Rep. 7: V108. 
 
Lizotte, E., and Miles, T.D. (2020) Hop disease management in Michigan for 2020. MSU 
Ext., https://www.canr.msu.edu/news/managing-hop-downy-mildew-in-michigan 
 
Michigan Department of Agriculture & Rural Development (2018) Michigan Agriculture 
Facts & Figures. 
https://www.michigan.gov/documents/mdard/MI_Ag_Facts__Figures_474011_7.pdf 
 
Mitchell, M.N., Ocamb, C.M., Grünwald, N.J., Mancino, L.E., and Gent, D.H. (2011) 
Genetic and pathogenic relatedness of Pseudoperonospora cubensis and P. humuli. 
Phytopathology 101: 805–818 
 
Morinha, F., Magalhães, P. and Blanco, G. (2020), Different qPCR master mixes 
influence telomere primer binding within and between bird species. J Avian Biol, 51: 
doi:10.1111/jav.02352 
 
Naegele, R. P., Quesada-Ocampo, L. M., Kurjan, J. D., Saude, C., and Hausbeck, M. K. 
(2016) Regional and temporal population structure of Pseudoperonospora cubensis in 
Michigan and Ontario. Phytopathology 106: 372–379 
 
NOAA (2019) National Climate Report - May to August 2019. 
https://www.ncdc.noaa.gov/sotc/national/201905 

 

74 

 

 

 
Ojiambo, P.S., Gent, D.H., Quesada-Ocampo, L.M., Hausbeck, M.K., and Holmes, G.J. 
(2015) Epidemiology and Population Biology of Pseudoperonospora cubensis: A Model 
System for Management of Downy Mildews. Annu Rev Phytopathol 53: 223–246 
 
Ojiambo, P.S., and Holmes, G.J. (2010) Spatiotemporal spread of cucurbit downy 
mildew in the Eastern United States. Phytopathology 101: 451–461 
 
Perla, D., Hausbeck, M.K., Cook, A. (2019) Evaluation of fungicide programs for control 
of downy mildew of cucumber, 2018. Plant Dis Manag Reports 147: V13 
 
Quesada-Ocampo L.M., Granke, L.L., Olsen, J., Gutting, H.C., Runge, F., Thines, M., 
Lebeda, A., and Hausbeck, M.K. (2012) The genetic structure of Pseudoperonospora 
cubensis populations. Plant Dis 96: 1459–1470 
 
Rahman, A., Standish, J., D’Arcangelo, K.N., Quesada-Ocampo, L.M. (2020) Clade-
specific biosurveillance of Pseudoperonospora cubensis using spore traps for precision 
disease management of cucurbit downy mildew. Phytopathology. doi: 10.1094/PHYTO-
06-20-0231-R 
 
Reuveni, M., Eyal, H., Cohen, Y. (1980) Development of resistance to metalaxyl 
in Pseudoperonospora cubensis. Plant Dis 64: 1108–1109 
 
Rogers, S.L., Atkins, S.D., and West, J.S. (2009) Detection and quantification of 
airborne inoculum of Sclerotinia sclerotiorum using quantitative PCR. Plant Pathol 58: 
324–331 
 
Royle, D.J. (1973) Quantitative relationships between infection by the hop downy 
mildew pathogen, Pseudoperonospora hutnuli, and weather and inoculum factors. Ann 
Appl Biol 73: 19–30 
 
Runge, F., and Thines, M. (2011) Reevaluation of host specificity of the closely related 
species Pseudoperonospora humuli and P. cubensis. Plant Dis 96: 55–61 
 
Salcedo, A., Hausbeck, M., Pigg, S., Quesada-Ocampo, L.M. (2020) Diagnostic Guide 
for Cucurbit Downy Mildew. Plant Heal Prog 21: 166–172 
 
Summers, C., Adair, N., Gent, D., Mcgrath, M., and Smart, C. (2015) 
Pseudoperonospora cubensis and P. humuli detection using species-specific probes 
and high definition melt curve analysis. Can J Plant Pathol 37: 150605144651006 
 
Thomas, A., Carbone, I., Choe, K., Quesada-Ocampo, L.M., and Ojiambo, P.S. (2017) 
Resurgence of cucurbit downy mildew in the United States: Insights from comparative 
genomic analysis of Pseudoperonospora cubensis. Ecol Evol 7: 6231–6246 
 
Thomas, A., Neufeld, K.N., Seebold, K.W., Braun, C.A., Schwarz, M.R., and Ojiambo, 
P.S. (2018) Resistance to Fluopicolide and Propamocarb and baseline sensitivity to 

 

75 

 

 

Ethaboxam among isolates of Pseudoperonospora cubensis from the Eastern United 
States. Plant Dis 102: 1619–1626 
 
Thornton, A.C, Adams, M.L., and Holmes, G.J. (2006) Evaluation of fungicides for 
control of downy mildew of cucumber I, 2006. Plant Dis Manag Rep 1: V122 
 
USDA (2020) United States Department of Agriculture. 2019. Vegetables Annual 
Summary. https://usda-library-cornell-
edu.proxy1.cl.msu.edu/concern/publications/02870v86p?locale=en 
 
Withers, S., Gongora-Castillo, E., Gent, D., Thomas, A., Ojiambo, P.S., and Quesada- 
Ocampo, L.M. (2016) Using Next-Generation sequencing to develop molecular 
diagnostics for Pseudoperonospora cubensis, the cucurbit downy mildew pathogen. 
Phytopathology 106: 1105–1116 
 
Zhu, S.S., Liu, X.L., Wang, Y., Wu, X.H., Liu, P.F., Li, J.Q., Yuan, S.K., and Si, N.G. 
(2007) Resistance of Pseudoperonospora cubensis to flumorph on cucumber in plastic 
houses. Plant Pathol 56: 967–975 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

76 

 

 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

 
 
 
 

 

 

 

 

 

 

 

 

 

CHAPTER III: OPTIMIZING SPORE TRAPS AND QUANTITATIVE PCR ASSAYS 

FOR THE MONITORING OF CUCURBIT DOWNY MILDEW. 

77 

 

 

ABSTRACT 
 

Current management of Pseudoperonospora cubensis, the causal agent of 

cucurbit downy mildew (CDM), relies on an intensive fungicide program. In Michigan, 

CDM occurs annually due to an influx of airborne sporangia; timely alerts of airborne 

inoculum can assist growers in assessing the need to initiate fungicide sprays. The 

main objective of our research was to improve the detection and quantification of 

airborne concentrations of P. cubensis sporangia by adapting two qPCR-based assays 

to distinguish between P. cubensis clade I and II and P. humuli in spore trap samples. 

We also aim to evaluate the efficiency of Burkard and impaction spore traps for the 

detection airborne concentrations of P. cubensis sporangia. A new qPCR assay 

improved the specificity of P. cubensis detection and resulted in a better linear 

correlation between the number of sporangia observed using light microscopy and Cq 

values obtained from Burkard spore traps (R2=0.6; p = 0.01). After two years of 

monitoring, P. cubensis clade II and P. humuli DNA were detected in air samples 

collected in commercial cucumber fields, while P. cubensis clade I DNA was not 

detected. P. cubensis clade II DNA was detected in spore trap samples >2 days before 

CDM symptoms were first observed in cucumber fields (August), while P. humuli DNA 

was only detected early in the growing season (May and June). P. cubensis clade I 

DNA was not detected in air samples before or after the disease onset in cucumber 

fields. Additionally, the probability for P. cubensis detection in Burkard spore trap 

samples was higher compared to impaction spore trap samples with approximately the 

same number of sporangia, suggesting that the efficiency of recovery of sporangia by 

Burkard spore traps exceeds the recovery of impaction spore traps. Our study identified 

 

78 

 

 

an improved methodology to monitor the airborne concentrations of 

Pseudoperonospora spp. sporangia using spore traps coupled with qPCR. This 

methodology could be used as part of a CDM risk advisory system to time fungicide 

applications that protect cucurbit crops in Michigan. 

INTRODUCTION 
 

Cucurbit downy mildew (CDM), caused by the obligate oomycete 

Pseudoperonospora cubensis, incites foliar blighting of several Cucurbitaceae species 

worldwide (Mitchell et al., 2011). Symptoms include irregular to angular chlorotic leaf 

lesions and pathogen sporulation on the lower leaf surface (Cohen et al., 2015) leading 

to premature defoliation, stunted plants and reduced yield, especially in cucumber 

(Reuveni et al., 1980; Adams et al., 2019; Hausbeck et al., 2019; Perla et al., 2019). In 

the U.S., P. cubensis overwinters on living hosts in regions that do not experience a 

frost or in heated greenhouses (Ojiambo et al., 2011; Naegele et al., 2016). The 

pathogen’s airborne sporangia disperse to northern U.S. growing regions annually from 

overwintering sources (Ojiambo & Holmes, 2010; Ojiambo et al., 2015). 

Michigan is the number one producer of pickling cucumbers and squash in the 

U.S. (USDA, 2020), but since 2005 CDM has threatened cucumber production annually 

in the state. In the absence of CDM resistant cucumber cultivars, growers rely on 

intensive fungicide spray programs to limit disease at a significant cost (Savory et al., 

2011), however, P. cubensis has developed resistance to key fungicides (Zhu et al., 

2007; Blum et al., 2011; Holmes et al., 2014; Keinath, 2015). For more than 10 years, 

information on the sporangial concentrations of P. cubensis derived from Burkard spore 

 

79 

 

 

traps has been used in Michigan as a decision-making tool to initiate fungicide sprays to 

control CDM (Granke and Hausbeck, 2011; Granke et al., 2013).  

Burkard and impaction spore traps are the most widely used devices for 

atmospheric sampling in plant pathology (Frenz, 1999). The Burkard spore trap has 

been used in aerobiological studies for over 60 years (West and Kimber, 2015) and 

operates by drawing air into a collection chamber containing a reel mounted onto a 

clockwork mechanism. Spores and other air-borne particles are impacted onto a 

greased tape covering the reel (Burkard Manufacturing Co. Ltd., U.K.) that rotates past 

the intake orifice at 2 mm/hr for 7 days. The impaction spore trap, also known as a 

rotating-arm spore sampler or rotorod spore trap, has become popular for the early 

detection of airborne plant pathogens (Jackson and Bayliss, 2011; Klosterman et al., 

2014; Fall et al., 2015; Choudhury et al., 2016a; Kunjeti et al., 2016). This device 

includes rods coated with adhesive material which spin at a standard rate in a rotating 

arm, impacting and collecting airborne particles (TSE Systems, Chesterfield, MO). 

However, frequent monitoring is required to obtain accurate estimations of inoculum per 

unit of time (i.e. an hour or day). Although Burkard spore traps are robust and highly 

autonomous, impaction spore traps can be more cost-effective and grower friendly (i.e. 

easier to use) (Jackson and Bayliss, 2011; Choudhury et al., 2016a). The efficiency of 

impaction spore traps to monitor P. cubensis sporangia in comparison to Burkard spore 

traps has not yet been assessed under field conditions.   

While the spore traps provide quantitative data on airborne sporangial 

concentrations, the processing and microscopic examination of the tapes and/or rods is 

time consuming and can result in misidentification due to the morphological similarities 

between species (Dung et al., 2018). A qPCR assay was designed for molecular 

 

80 

 

 

differentiation of the morphologically identical sporangia of P. cubensis and 

Pseudoperonospora humuli, the causal agent of hop downy mildew (HDM) (Summers et 

al., 2015a). The detection of each species is based on the recognition of a conserved 

single nucleotide polymorphism (SNP) in the cytochrome oxidase subunit II (cox2) 

gene. However, quantification of sporangia using this assay was compromised in spore 

trap samples containing DNA from both species due to the high similarities between 

them in the region targeted (Bello et al., 2020).  

To further improve the specificity and facilitate the quantification of P. cubensis 

and P. humuli sporangia, a new qPCR assay targeting unique sequences in the 

pathogens’ mitochondrial genome was developed that also enables detection and 

differentiation of both species in a single reaction (Crandall, 2020). This assay can also 

differentiate between P. cubensis clades I and II. Isolates belonging to these two clades 

are host-adapted at the cucurbit species level (Summers et al., 2015b; Thomas et al., 

2017a; Wallace et al., 2020). Clade I isolates are recovered more frequently from 

commercial varieties of Cucurbita pepo, C. moschata, C. maxima, and Citrullus lanatus 

and clade II isolates are associated more frequently with commercial varieties within the 

Cucumis genus (Wallace et al., 2020). Accurate identification of airborne sporangia of 

P. humuli and P. cubensis (clades I and II) is critical to monitor the pathogen’s arrival 

into Michigan’s growing regions so as to inform fungicide applications. This new 

multiplex qPCR assay has the potential to improve quantification of airborne sporangia 

in spore trap samples without sacrificing accuracy. 

The main objective of our research was to improve the detection and 

quantification of airborne concentrations of Pseudoperonospora spp. sporangia by 

adapting two qPCR-based assays to distinguish between P. cubensis and P. humuli in 

 

81 

 

 

spore trap samples collected in the field. The specific detection and monitoring of 

sporangia from each clade of P. cubensis (clade I and II) in commercial cucumber fields 

was of particular interest. Additionally, we also aimed to evaluate Burkard and impaction 

spore traps for their efficiency in detecting airborne concentrations of P. cubensis 

sporangia. 

MATERIALS AND METHODS 
 

Pseudoperonospora isolates. To perform in vitro evaluations, a single-lesion 

isolate of P. cubensis clade II (CDM23 cucumber) and clade I (CDM-YUM squash), and 

a single-spike isolate of P. humuli (isolate HDM19) were maintained as described by 

Bello et al. (2020). Sporangia from each isolate were rinsed from the host tissue into 

falcon tube with distilled water and the sporangial suspensions were concentrated by 

centrifugation (5424R centrifuge, Eppendorf, New York, NY) at 14.000 rpm for 5 min. 

The resulting pellet was homogenized in impact-resistant 2mL tubes (Lysing Matrix H, 

MP Biomedicals, Irvine, CA) using a TissueLyser II (Qiagen, Valencia, CA) for 4 min at 

30 Hz. DNA was extracted using a NucleoSpin Plant II isolation kit (Macherey-Nagel, 

Bethlehem, PA) following manufacturer’s instructions and the DNA concentration was 

determined using the Qubit double-stranded DNA High Sensitivity Assay Kit (Life 

Technologies, Carlsbad, CA). 

Multiplexing of qPCR assays. All qPCR experiments were conducted using the 

protocols described by Bello et al. (2020) and Crandall et al. (2020) referred to hereafter 

as protocol A and B, respectively. All qPCR reactions were manually assembled into 96-

well white plates (Bio-rad MLL9651) containing 10 µl of the Prime-Time Gene 

Expression Master Mix (IDT, Skokie, IL), 2µl of template DNA, and 8 µl of a solution 

 

82 

 

 

containing  primers, probes and internal controls as described in Table 3-1. Negative 

control reactions lacking a DNA template were included in each plate run. The qPCR 

protocols were run on a CFX 96 Touch qPCR system (Bio-rad); the cycling conditions 

are summarized in Table 3-1. Two technical replicates of each sample were run and the 

average Cq and standard deviation were calculated using Bio-rad CFX Manager 

software (version 3.1).  

Specificity and sensitivity of qPCR assays. The sensitivity and specificity of 

the qPCR protocols were tested using ten-fold dilutions of genomic DNA from two 

independent DNA extractions of each isolate (CDM23, CDM-YUM and HDM19). Three 

technical replicates of each sample dilution were tested using both qPCR protocols 

(Table 3-1) and the average Cq values with standard deviation were calculated using 

Bio-rad CFX Manager software. Mean Cq values were plotted against the log10 of 

template DNA concentrations and used to generate standard curves. To assess the 

specificity of the qPCR protocol B, samples of mixed DNA from the three isolates were 

evaluated to determine whether the assay could detect P. cubensis (clades 1 and 2) 

and P. humuli. Ten-fold dilutions of genomic DNA from each isolate were mixed in 

varying concentrations (Table 3-2) and subjected to qPCR. Three technical replicates of 

each mixture were run and the average Cq and standard deviations were calculated 

using Bio-rad CFX Manager software. 

To evaluate the relationship between sporangial concentrations and the Cq 

values from the multiplexed qPCR assays in vitro, dilution series containing 1, 3, 5, 10, 

25 and 50 sporangia were prepared as described by Crandall et al. (2020) and 

regressed against the corresponding Cq values of each assay. There were fifteen 

replicates each for the sporangial counts of 1, 3, and 5 and eight replicates each for 

 

83 

 

 

sporangial counts of 10, 25, and 50. DNA was extracted using a NucleoSpin Plant II 

isolation kit (Macherey-Nagel, Bethlehem, PA) following manufacturer’s instructions. 

Collection and qPCR evaluation of field samples. Airborne sporangial 

concentrations were monitored during the growing season (May to August) in 2018 and 

2019 using Burkard and impaction spore traps. Each year, a Burkard spore trap and an 

impaction spore trap were placed side by side approximately 20 m from commercial 

cucumber fields in the Michigan counties of Muskegon, Allegan, Bay, and Saginaw (Fig. 

3-1). In 2019, a cucumber research plot (0.25 ha) at the Michigan State University 

(MSU) Plant Pathology Farm located in Ingham County was also monitored using a 

Burkard spore trap and impaction spore trap. An abandoned hop research yard (0.25 

ha) with basal shoots infected by P. humuli was located 200 m from the cucumber 

research plot. A Burkard spore trap, only, was placed in a commercial hop yard in 

Berrien County in 2019 (Fig. 3-1).  

Burkard spore traps were set to an approximate airflow rate of 10 l/min. The reel 

of each spore trap was covered with a melinex tape coated with an adhesive mixture of 

petroleum jelly and paraffin (9:1 wt/wt) dissolved in sufficient toluene to provide the 

desired thickness. The tape was removed weekly and cut longitudinally along the center 

line in two subsection of 9 x 336 mm each (Rogers et al., 2009). The two resulting sub-

section were processed as described by Bello et al. (2020). Briefly, the first section was 

cut at 24-hr segments (48-mm lengths) and subjected to DNA extraction. The second 

section was also cut at 24-h segments and screened using light microscopy to estimate 

the number of sporangia captured per day (Granke et al., 2013). 

The impaction traps were constructed using a motor (RF-500TB-10750, 

Mabuchi) that spun at 2700 rpm when powered by a 12-V marine battery. Each 

 

84 

 

 

impaction trap was mounted onto a 1.5-m tower that holds the collection rods 1 m 

above the ground. The impaction traps operated continuously throughout each growing 

season (May to August), and the greased-coated rods (1.2-by-3.5-mm stainless steel) 

were collected four times per week at intervals of 24, 24, 24, 96 h (Table 3-3) 

(Choudhury et al., 2016a). Samples from Burkard and impaction traps were carefully 

placed into impact-resistant 2mL tubes (Lysing Matrix H, MP Biomedicals) containing 

100 µl of PL1 buffer (Macherey-Nagel, Bethlehem, PA). The samples were 

homogenized using a TissueLyser II (Qiagen, Valencia, CA) for 4 min at 30 Hz and 

DNA was extracted using NucleoSpin Plant II isolation kit (Macherey-Nagel, Bethlehem, 

PA) following manufacturer’s instructions. Subsequently, 2 µl of the extraction products 

were evaluated using qPCR. All 2019 samples were assessed using both A and B 

protocols, however, all the samples collected during 2018 were evaluated using only 

protocol A.    

Correlating Cq values with sporangial counts of field samples. Linear 

regression using R (version 3.6.1) was used to assess the relationship between the Cq 

values of each qPCR assay and the number of sporangia in trap samples. The Cq 

values of each qPCR assay were regressed against the number of sporangia quantified 

using light microscopy in the second half of the tape of the Burkard spore traps. The 

linear equation obtained from the regression analysis (Table 3-4) was used to determine 

the corresponding number of sporangia expressed as NCq where N represents the 

number of sporangia calculated using Cq values.  

Additionally, binary logistic regression was used to model the relationship 

between the number of airborne sporangia and the probability of a positive detection in 

samples collected by Burkard or impaction spore traps. The log10 number of sporangia 

 

85 

 

 

quantified in the first half of the Burkard spore traps were regressed with the results 

from protocol A obtained from trap samples as a categorical (binary) variable. The Cq 

values of protocol A were categorized using a threshold Cq value of 35.5 (Cq>35.5 

negative and Cq≤35.5 positive). The regression was performed independently using the 

qPCR results from each spore trap separately (i.e. Burkard or impaction). We used this 

approach to compare between traps and estimate the probability of a positive detection 

of P. cubensis DNA given an approximate number of sporangia in the air. For example, 

the probability of a positive detection in impaction samples given an estimated number 

of 10 sporangia in air samples. Assuming that p is the probability of a positive detection 

by the qPCR, the logistic regression line is described by Equation 1.  

 

Equation 1: Binary logistic regression used to model the relationship 
between the number of airborne sporangia and the probability of a positive 
detection in spore trap samples. 

"#$!"% &'−	&*	=	b"+ b!∗"#$!"(/!) 
 & is the probability of a positive detection. b0 to b1 are parameters.  

 

 

RESULTS 
 

Sensitivity and specificity of qPCR assays. Using ten-fold dilutions of genomic 

DNA, the probes of both protocols detected P. humuli and each clade of P. cubensis 

within total DNA template amounts between 100 to 1000 fg (R2³ 0.99; p < 0.05) (Fig. 3-

2 and 3-3). The assays are intended to be used to quantify sporangia in environmental 

samples, thus, further testing outside of these template amounts was not done. Both 

protocols detected 100 fg of each taxon with average Cq values below 37.5 as reported 

 

86 

 

 

previously (Bello et al., 2020; Crandall et al., 2020; Summers et al. 2015). Most samples 

with concentrations below 100 fg were not detected reliably (>95% of the times tested), 

thus, 100 fg of template DNA was considered as the lower limit of detection (LOD) of 

the qPCR assays for both species (Figs. 3-2 and 3-3).  

In singleplex reactions with DNA recovered from isolated sporangia of each 

Pseudoperonospora taxa, unspecific amplification was observed using protocol A (Fig. 

3-2 A). On the contrary, singleplex reactions of protocol B using DNA recovered from 

isolated sporangia allowed the specific detection of each taxon and background 

amplification of nontarget taxa was not observed (Fig. 3-3 A, B, C). Multiplex reactions 

containing all the primers and probes of each qPCR protocol yielded the same results.  

Small, yet significant differences in the Cq values of mixed and unmixed DNA 

samples assessed using protocol A were previously reported (Bello et al., 2020). Similar 

results were also observed using protocol B (Table 3-2). The presence of DNA from a 

second non-target taxa slightly affected the sensitivity of the probes Pcub2 and 

Pcub_RFLP_qP1 designed for the detection of P. cubensis clades II and I, respectively 

(Table 3-1). When 100 fg of P. cubensis clade II DNA were mixed in a 1:1 or 1:100 ratio 

with DNA from a second taxon, no amplification or a significant increase in the Cq value 

was observed with the Pcub2 probe. Similarly, when 100 fg of P. cubensis clade I DNA 

were mixed in a 1:100 ratio with DNA from a second taxon, no amplification was 

observed with the probe Pcub_RFLP_qP1 (Table 3-2). 

The regression between sporangial concentrations and the Cq values of both 

assays (A and B) demonstrated reliable pathogen detection with as few as three 

sporangia for each taxon (Fig. 3-11 and 3-12). Amplification results of DNA from a 

 

87 

 

 

single sporangium were inconsistent and outside of the linear relationship between 

sporangial counts and Cq values. 

Correlation between Cq values and sporangial counts of Burkard spore 

trap samples. DNA of P. cubensis clade II and P. humuli was detected using both 

qPCR protocols on field samples containing fewer than 10 sporangia (Fig. 3-4). 

Regression analysis indicated that the number of sporangia on the second half of the 

tape of field samples (quantified using light microscopy) explained 37% and 60% of the 

variation in the Cq values of P. cubensis clade II DNA obtained using protocols A and B, 

respectively (Fig. 3-4 A). Significant differences between the protocols were detected 

using regression analysis (p<2.24e-05, Table 3-4) with a more inclined regression line 

observed for protocol B (Fig. 3-4 A). Similarly, the number of sporangia on the second 

half of the tape explained 27% and 30% of the variation in the Cq values of P. humuli 

DNA obtained using protocols A and B, respectively (Fig. 3-4 B). No significant 

differences between the protocols were detected (P=0.26, Table 3-4) using regression 

analysis. b estimates of equations 2 and 3 that describe the relationship between the Cq 

values of each qPCR assay and the sporangial numbers of each species are 

summarized in Table 3-4.  

 

Equation 2: Regression line that describes the relationship between Cq 

values and sporangial numbers of Pseudoperonospora cubensis clade II. 

 12#.%&'	)= b"− b!	"#$!"(3&#456$75)+ b)	"#$!"(2819)b*	"#$!"(3&#456$75)(2819	) 

 
The values of the b0, b1, b2, b3 parameters are summarized in Table 3-4. The qPCR 
assay is express as a binary variable: qPCR assay A = 1, qPCR assay B = 0. 
 
 
 

 

88 

 

 

Equation 3: Regression line that describes the relationship between Cq 

values and sporangial numbers of Pseudoperonospora humuli. 
 
 

12#.+&= b"− b!	"#$!"(3&#456$75)). 

 
The values of the b0 to b1 parameters are summarized in Table 3-4. 
 
 

Logistic regression. After categorizing the qPCR results of both spore traps 

(positive/negative) a higher number of impaction samples were negative compared to 

the Burkard samples (Fig. 3-5 C, D). This occurred most frequently when the 

atmospheric concentration of sporangia was below 100 sporangia/day as estimated 

using microscopic analysis of the Burkard tape. Using a logistic regression to model the 

relationship between the number of atmospheric sporangia and the probability of 

detection, higher estimates were obtained for the Burkard traps (Table 3-5). Therefore, 

a higher probability of P. cubensis detection given any number of sporangia was 

estimated for Burkard trap samples (Figs. 3-5 A, B). A probability of detection above 

90% was obtained for Burkard and impaction trap samples with an approximate number 

of sporangia equal to 15 and 120 sporangia, respectively.     

Assessment of field samples using light microscopy and qPCR. During 

2018 and 2019, an average of 15 reels of Burkard spore traps were collected from May 

to August in all the locations monitored (Table 3-3, Figs. 3-5 to 3-10). The tape from 

each reel was divided by days for microscopic and qPCR analysis generating a total of 

105 to 112 samples of 24 h per location each year (Table 3-3, Figs. 3-5 to 3-10). 

Similarly, during both years a total of 544 impaction trap samples were collected from 

May to August among all the locations monitored. An average of 45 and 15 impaction 

 

89 

 

 

trap samples per location were collected every 24 and 96 h, respectively (Table 3-3, 

Figs. 3-5 to 3-10). The qPCR assays detected Pseudoperonospora spp. DNA on 

41.84% of the 24 h-samples collected using Burkard spore traps (Table 3-3). Similarly, 

detection of Pseudoperonospora spp. DNA occurred in 23.35 and 29.81% of the 

impaction trap samples collected every 24 and 96h, respectively (Table 3-3).   

 

In 2018, Pseudoperonospora spp. sporangia were first observed in May or June 

across the monitored cucumber fields (Figs. 3-6 and 3-7). During May, June and July, 

the number of sporangia observed in the tape of Burkard spore traps remained below 

10 sporangia/day and P. cubensis clade II DNA was detected with Cq values above 31 

(Fig. 3-6 and 3-7). This Cq value corresponds to <3Cq sporangia according to the 

regression line that describes the relationship between Cq values and sporangial 

numbers (Equation 2 and Table 3-4). Using light microscopy, 10 to 40 sporangia/day 

were detected in June in Bay Co. (22, 27 June) and Allegan Co. (6, 7, 24 June) and 

were identified as P. humuli using qPCR with Cq values between 31 and 27 

(corresponding to 3Cq to 15Cq sporangia) (Figs. 3-6 B and 3-7 A). In all locations, the 

number of airborne Pseudoperonospora spp. sporangia observed in the tape of the 

Burkard tape increased during the last week of July and reached a maximum in the 

second or third week of August. During this time, P. cubensis clade II DNA was 

regularly detected in all fields with Cq values that range between 31 and 22 

(corresponding to 3Cq and 1200 Cq sporangia) but P. humuli was not detected (Figs. 3-6 

and 3-7). In 2018, CDM symptoms were detected in August for each monitored 

cucumber field, after 12 to 40 sporangia/day were observed in the Burkard spore trap 

tape. Correspondingly, before symptoms were observed in the fields, P. cubensis clade 

 

90 

 

 

II detection using qPCR occurred one to seven days earlier with Cq values below 31 

(equivalent to >3Cq sporangia) (Figs. 3-6 and 3-7).  

Similar results were also obtained after the evaluation of impaction trap samples 

using qPCR in 2018. However, when the number of sporangia per day was low (<20 

sporangia/day), pathogen detection in impaction trap samples using qPCR was less 

consistent. For instance, 1 to 20 P. humuli sporangia were observed in the first half of 

the Burkard spore trap tape in June (Bay and Allegan Counties) and detected in the 

second half using qPCR, but P. humuli detection did not occur in the impaction trap 

samples of the same dates (Figs. 3-6B and 3-7A). Similarly, before CDM symptoms 

were observed, P. cubensis DNA was consistently detected in Burkard samples in all 

the fields monitored, but it was only detected in impaction trap samples in Muskegon 

and Allegan Counties. (Figs. 3-6 A and 3-7 A). Using the Burkard spore trap, P. 

cubensis clade II DNA was detected with Cq values below 31 approximately 10 days 

before CDM symptoms were observed in the field (Figs. 3-6 and 3-7). On the other 

hand, P. cubensis clade II DNA was detected in impaction trap samples approximately 7 

days before symptoms were observed in the field only in Muskegon and Allegan 

Counties. (Figs. 3-6 and  3-7, Table 3-6). 

In 2019, the number of sporangia observed in Burkard samples from May to July 

were generally below 10 sporangia/day, but exceptions occurred in May and/or June in 

Muskegon, Allegan, Saginaw, Ingham and Berrien Counties where daily counts reached 

values between 10 to 40 sporangia (Figs. 3-8, 3-9 and 3-10). From May to July, P. 

cubensis clade II DNA was occasionally detected with Cq values above 36 

(corresponding to <3Cq sporangia) in all fields monitored. When sporangial numbers 

were above 10 sporangia/day, the Cq values for P. humuli detection reached values 

 

91 

 

 

between 31 and 25. These Cq values correspond to >3Cq and 50 Cq sporangia according 

to the regression line that describes the relationship between Cq values and sporangial 

numbers (Equation 2 and Table 3-4). Most Cq values for P. humuli detection below 31 

were detected only during June, when the highest numbers of sporangia (>40) were 

observed using light microscopy. 

In the commercial cucumber fields monitored, the concentration of 

Pseudoperonospora spp. sporangia estimated using light microcopy increased during 

August but did not reach the numbers observed in 2018 (Figs. 3-8 and 3-9). At these 

locations, CDM symptoms were detected during the third or fourth week of August. 

More regular detection of P. cubensis clade II DNA in Burkard trap samples with Cq 

values between 31 and 20 (corresponding to 3Cq to 400Cq sporangia) occurred after the 

observation of CDM symptoms. P. cubensis detection using qPCR before CDM 

symptoms were observed in the cucumber fields occurred one two seven days earlier 

with Cq values between 36 and 30 (corresponding >3Cq to 10Cq sporangia) (Figs. 3-8 

and 3-9). On the other hand, P. humuli DNA was detected almost every day from May 

to August in the commercial hop yard (Fig. 3-10 B) and from May to July in the 

cucumber research plot (Fig. 3-10 A). At these two locations, concentrations between 

10 and 40 sporangia were observed from the Burkard trap tapes during May and June 

and Cq values between 31 and 23 (corresponding to >3Cq to 100Cq sporangia) were 

registered (Fig. 3-10). 

The monitoring of Pseudoperonospora spp. using impaction trap samples in the 

commercial cucumber fields had similar results to those of the Burkard spore traps in 

2019 (Fig. 3-8, 3-9 and 3-10). However, when atmospheric concentrations of P. 

cubensis sporangia were below 10 sporangia/day (estimated with the aid of a light 

 

92 

 

 

microscope), DNA detection occurred less frequently in samples collected by impaction 

spore traps compared to the Burkard spore traps. Using qPCR, P. cubensis was 

detected in samples from both traps approximately two weeks before symptoms were 

observed in the cucumber fields monitored in Muskegon, Allegan and Saginaw 

counties. Only in Bay county, P. cubensis was not detected two weeks before 

symptoms developed in impaction trap samples (Table 3-6). Generally, P. cubensis 

detection in 2019 occurred with higher Cq values compared to 2018. 

DISCUSSION 
 

 
Early detection and quantification of airborne P. cubensis sporangia could 

improve the timing of fungicide initiation in Michigan as the pathogen is reintroduced to 

northern U.S. production regions each year (Bello et al., 2020). The Burkard spore traps 

have been used in the state for this purpose since 2008, however, the inability to 

distinguish between the morphologically identical sporangia of P. humuli and P. 

cubensis has been a significant shortcoming. Using an improved qPCR assay (Crandall 

et al., 2020), we were able to distinguish between three host-adapted 

Pseudoperonospora taxa in spore trap samples: P. humuli, P. cubensis clade I, and P. 

cubensis clade II. During two years of monitoring using Burkard and impaction spore 

traps coupled with qPCR in commercial cucumber fields, P. cubensis clade II sporangia 

were detected 2 to 10 days before CDM symptoms were observed. Both spore traps 

recorded similar trends in the airborne concentration of P. humuli and P. cubensis 

(clade I and II) sporangia, however, in our study, the Burkard spore trap was a more 

efficient instrument for detecting P. cubensis sporangia.   

 

93 

 

 

The multiplex qPCR assay developed by Crandall et al. (2020) that targets the 

open reading frames (orf 374, orf 367, orf 329) in the mitochondrial genome of each 

taxon provided high specificity for detecting DNA of P. humuli and each clade of P. 

cubensis (Crandall et al., 2020). This qPCR assay allowed us to estimate the 

concentrations of P. cubensis and P. humuli sporangia in Burkard and impaction spore 

trap field samples. The new multiplex qPCR assay had the same sensitivity of the qPCR 

assay developed by Summers et al. (2015), which was also developed to target a 

mitochondrial DNA region. Both assays detected P. cubensis and P. humuli DNA at 

amounts ranging from 100-1000 fg in vitro and were equally sensitive to a qPCR assay 

targeting a nuclear DNA region (Rahman et al., 2020). This high sensitivity was 

validated with the detection of less than 10 sporangia of each species in field spore trap 

samples. In vitro, samples containing three sporangia of each clade were reliably 

detected (Crandall et al., 2020) using these qPCR assays. Comparatively, the qPCR 

assay that targets the single-copy nuclear gene c255.3e7 showed a detection limit of 10 

sporangia (Rahman et al., 2020). Additionally, the detection of each taxon without cross 

reactivity in samples containing mixed DNA of P. humuli and either clade of P. cubensis 

validated the high specificity of the assay developed by Crandall et al. (2020). This 

improved specificity was possible because the primers and probes of this new qPCR 

assay target more polymorphic regions in the mitochondrial genome compared to the 

single nucleotide polymorphism used for differentiation in the assay developed by 

Summer et al. (2015).  

Improving the specificity also resulted in an increased linear correlation between 

the number of P. cubensis sporangia quantified with the aid of a light microscope and 

the Cq values of the qPCR (R2=0.6). We used the equation that describes this 

 

94 

 

 

relationship to calculate the number of sporangia based on Cq values; however, the 

standard error of the b estimates used in this equation suggests a low precision for the 

quantification of sporangia using qPCR in spore trap samples. Multiple factors such as 

high variation during DNA extraction (Summers et al., 2015a), user variation during 

qPCR, and the multicopy nature of the target mitochondrial genes (González-

Domínguez et al., 2020) could explain these results. The number of mitochondria can 

vary greatly among cells (O’Hara et al., 2019) reducing the precision for the 

quantification of cells using qPCR assays that target mitochondrial genes, however, the 

additional number of mitochondria per cell can increase the ability to detect low 

concentrations of sporangia. The more precise the assay (i.e. regression model) the 

closer predictions are to the observed number of sporangia. In this study, the use of Cq 

values to predict the number of sporangia resulted in misestimates compared to the 

quantification using light microscopy. Single-copy nuclear genes are thought to offer 

more precision for cell quantification compared to mitochondrial genes. However, the 

correlation of Cq values obtained from the amplification of the single-copy nuclear gene 

c255.3e7 with sporangial numbers revealed a standard deviation across more than 

three amplification cycles for samples with the same number of sporangia (Rahman et 

al., 2020). Similar results were obtained using the mitochondrial genes orf 374, orf 

367and orf 329 (Crandall et al., 2020) suggesting that the precision for cell 

quantification using qPCR is not significantly increased with the utilization of nuclear 

genes as amplification targets.  

During the two years of monitoring in commercial cucumber fields in Michigan, P. 

humuli detection occurred early in the growing season. This was expected because hop 

downy mildew is prevalent in the state (Higgins et al., 2020; Lizotte et al., 2020) and the 

 

95 

 

 

pathogen overwinters in dormant hop crowns, growing into expanding basal shoots in 

the spring (Coley-Smith, 1962). Similarly, detection of P. cubensis clade II occurred as 

expected because airborne sporangia are dispersed every year from overwintering 

sources resulting in CDM outbreaks in cucumber production regions of the upper 

Midwest (Ojiambo et al., 2011; Naegele et al., 2016). However, P. cubensis clade I was 

not detected in the 980 Burkard and 544 impaction spore trap samples that were 

collected over the two years of sampling from May to August. The absence of P. 

cubensis clade I sporangia in the air samples collected at the monitored cucumber fields 

may be due to the reduced number of crops planted in the state that are known to be 

hosts of this clade (i.e. C. pepo, C. moschata, C. maxima, and C. lanatus) (Wallace et 

al., 2020). Approximately 6000 ha of C. maxima (pumpkin), C. pepo (squash). C. 

moschata (butternut squash) and C. lanatus (watermelon) are planted in Michigan, 

compared to the more than 15000 ha of C. sativus (cucumber) planted in the state 

every year (USDA, 2020).  

In North Carolina, the hectares of cucumber planted annually are also greater 

compared than those planted with other cucurbits; clade I was consistently detected 

using spore traps in the fall (September and October), while clade II was detected in the 

summer and fall (Rahman et al., 2020). In our study, we report on the airborne 

concentration of P. cubensis sporangia through August and most of the cucumber 

acreage is harvested by this time. In Michigan, the foliage of pumpkins and hard squash 

during September and October begins to senesce and is compromised by powdery 

mildew; CDM is rarely reported on these crops in the state. Determining if sporangia 

from both clades are dispersed to the Great Lakes cucurbit growing regions during 

September to October remains to be seen by future studies. P. cubensis clade I was not 

 

96 

 

 

detected in North Carolina and Michigan during the summer probably due to the total 

area planted to hosts susceptible to this clade. In the absence of airborne P. cubensis 

clade I sporangia, an intensive fungicide program for CDM may be unnecessary in non-

cucumber hosts. Timely regional information on the atmospheric concentrations of each 

clade of P. cubensis could inform control measures to minimize the negative impact of 

CDM across different cucurbits.  

This information could be used as a decision-making tool to initiate fungicide 

sprays to protect susceptible crops, as it is used in other crop production systems  

(Carisse et al., 2009; Fall et al., 2015; Dhar et al., 2019; Van der Heyden et al., 2020). 

In lettuce, a threshold Cq value of 24, equivalent to 324 Bremia lactucae sporangia/day, 

is used to determine whether fungicides should be applied. This approach reduced the 

number of fungicide applications to control downy mildew in small lettuce plots without a 

significant increase in disease incidence (Dhar et al., 2019). Similarly, in our study, 

CDM symptoms were observed after airborne sporangial concentrations exceeded 10-

15 sporangia/day (estimated using light microscopy) or Cq values between 33 to 30 

(qPCR assay B) were detected in Burkard spore trap samples. This suggests that the 

critical concentration to trigger fungicide sprays against CDM could be close to these 

numbers. However, further research is required to better understand the interaction 

between sporangial concentrations, environmental conditions, and symptom 

development (Fall et al., 2015), which could ultimately lead to the establishment of 

spore concentration thresholds to trigger fungicide application. 

Accurate sampling of low inoculum loads and real-time monitoring is critical to 

develop a biosurveillance system that accurately assesses the risk of CDM in cucurbits. 

Burkard and impaction spore traps are the most widely used devices for atmospheric 

 

97 

 

 

sampling in plant pathology (Frenz, 1999) and have played a significant role in 

epidemiological studies in horticultural and agricultural settings (Granke et al., 2013; 

Choudhury et al., 2016a; Carisse et al., 2017; Wyka et al., 2017). They have also been 

used to accelerate the detection of airborne plant pathogens (Jackson and Bayliss, 

2011) prior to symptom development (Villari et al., 2016; Thiessen et al., 2017; Dung et 

al., 2018; Dhar et al., 2019). However, Burkard and impaction spore traps had not been 

used side by side to monitor Pseudoperonospora spp. sporangia. After two years of 

monitoring, our results suggest that the Burkard spore traps are a more efficient device 

detecting airborne sporangia at low concentrations (<100 sporangia/day). This is 

consistent with theoretical expectations that impaction spore traps are likely to offer 

lower particle recoveries than Burkard spore traps (Frenz, 1999).  

We found that the probability of P. cubensis detection is above 90% for Burkard 

and impaction samples that contain approximately 15 and 120 sporangia, respectively. 

These results indicate that Burkard spore traps can collect at least eight times (120/15) 

more sporangia than impaction spore traps at the conditions tested. The difference in 

the spore recovery between these two spore traps is expected to increase inversely 

proportional to the size of the particles being collected. According to Aylor (1993), 

Burkard spore traps can recover three times more particles than impaction spore traps 

when the particle size is approximately the same as that of P. cubensis sporangia (40 

µm). However, differences of up to seven times in particle recovery between these two 

collection devices have also been reported (Solomon et al., 1980).  

Wind direction and wind velocity can also affect the particle collection efficiency 

of both spore traps (West and Kimber, 2015), but the more important factors in the 

performance of impaction spore traps are sampling surface width and angular velocity 

 

98 

 

 

(Solomon et al., 1980). We used rods with similar width surface (1.2-by-3.5-mm 

stainless steel) to rods utilized in previous studies (Klosterman et al., 2014; Rahman et 

al., 2020) but increasing the collection surface of the impaction rod samplers could 

result in higher particle recovery and should be considered for future monitoring studies. 

Other factors to consider for increasing the recovery of P. cubensis sporangia by 

impaction spore traps include using multiple traps per location and longer deployment 

times of spore traps before rod collection. In California, at least two impaction spore 

traps are utilized to monitor B. lactucae spore loads in lettuce fields 50 to 200 times 

smaller than the commercial fields monitored in this study (Dhar et al., 2019); doubling 

the sample surface and the amount of air sampled could increase the chances for 

impaction of airborne sporangia in daily samples. We also observed a higher proportion 

of positive samples among the impaction spore trap samples collected over a longer 

period of time (96 h) which indicates that the adhesive medium we used on the rods 

(High-Vacuum Grease) was not completely saturated. It may not be necessary to 

change the impaction rods every 24 to 72 hours. Considering the expense of Burkard 

spore traps and the relatively low cost of impaction spore traps and their ease of use, it 

is important to improve the efficiency of sporangia detection using impaction spore 

traps. 

The combination of Burkard and impaction spore traps with the qPCR assay 

developed by Crandall et al., 2020 facilitated the sensitive and specific monitoring of P. 

humuli and two host-adapted clades P. cubensis in Michigan. Using spore traps and 

qPCR, we detected P. cubensis clade II sporangia three to seven days before disease 

onset in commercial cucumber fields. During two years of monitoring,  P. cubensis clade 

I was never detected during the summer season (May to August) in the fields 

 

99 

 

 

monitored. Our data suggest that Burkard spore traps are more efficient than impaction 

traps for the detection of airborne Pseudoperonospora spp. sporangia at low 

concentrations (<100 sporangia/day). Impaction spore traps could be modified to 

increase the probability for the collection of sporangia. In the future, the ability to rapidly 

detect both clades of P. cubensis using qPCR could be incorporated with environmental 

data and disease development information as part of a CDM risk advisory system to 

time fungicide applications that protect cucurbit crops in Michigan. 

 

100 

 

 

 
 
 
 
 
 
 
 
 
 
 

 
 
 
 
 
 
 
 
 
 
 

 

 
 
 
 

 

APPENDIX 

101 

 

 

APPENDIX 

 
Table 3-1. qPCR assays designed for the differentiation of Pseudoperonospora humuli and Pseudoperonospora 
cubensis clade I and II. 

Assay  Primers/probes 

Final 
concentr
ationa  
60 µm  
RT33F 
RT182R 
60 µm  
CUBprobeSNP105  2.5 µm 
HUMprobeSNP105  5 µm 
InCp_J2 
5 µm 

Internal control (IC)  7.5 x10-10 

nM 

PC_RFLP_2F 
PC_RFLP_3R 
PH_RFLP_4R 
PC-4 F 
PC-4 R 
Pcub_RFLP_qP1 

Pcub2 

20 µm 
10 µm 
10 µm 
10 µm 
10 µm 
5 µm 

10 µm 

Phum_RFLP_qP4 

5 µm 

Sequence 5'->3' 

Protocolb 

AACTCCCGTTATGGAAGGTATT 
CCATGTACAACAGTAGCTGGA 
HEX/A+C+AAA+C+G+AATA+CT/BHQ 
FAM/AA+C+AAA+C+A+AATA+CTG/BHQ 
CYS/A+GCATTATT+GTTTAT+CATATATACA/BHQ 
AACTCCCGTTATGGAAGGTATTATCATTAATCAT
GATTTGTA 
TATATGATAAACAATAATGCTATAACATAGAGTC
TCTTTCAT 
GAATAATCCAGCTACTGTTGTACATGG. 
CTGCTTTATCTTTTTCTTTTTG  
AGAGAAGATTTAGATTATAATTC  
AGAGACGATTTGGATTATAATT  
CAAGACCACCATTTTTATGTC  
TGGAAATTAAAAATTTTCTATTAC  
FAM/AACAAACTCAAGTAGAACTTCAACAAA/BH
Q 
HEX/AGGATTGATTTTCATTAATTCCTTTTTGTAA
TAGAA/BHQ 
RED/CCAACAGTTATACTTGTAATAAAC 
ATCAAG/BHQ 

 95°C for 3 min 
followed by 40 
cycles of 95°C 
for 10 s and 
65°C for 45 s 

95⁰C for 3 
minutes 
followed by 45 
cycles of 95⁰C 
for 15 s and 
58⁰C for 45 s 

Ac 

Bd 

 

 

102 

 

Table 3-1. (cont’d) 
 
a. Final concentrations used in a 20 ul qPCR reaction.  
b. The amplification protocols run on a CFX 96 Touch qPCR system (Bio-rad).  
c. This set of primers and probes were adapted from (Summers et al., 2015a). Locked nucleic acids in the probes are 
followed by a plus (+) sign.  
d. This set of primers and probes were adapted from (Crandall et al., 2020). 

 

 

103 

 

Table 3-2. Threshold cycle (Cq) values of the qPCR protocol B using varying concentrations of genomic DNA 
from Pseudoperonospora cubensis clade 1 and 2 and P. humuli  

Species/Clade DNA 

Probe 

Mix 
ratio 

Un-
mixed 

1/1 

 

Pcub2 

Cq±SD 

Phum_ 
RFLP_qP4 
Cq±SD 
NA 
NA 
NA 
28.17±0.47a  NA 
32.01±1.1b  NA 
34.11±0.25c  NA 
NA 
NA 
NA 

Pcub_ 
RFLP_qP1 
Cq±SD 
27.53±0.35a   NA 
-- 
31.69±0.21b  NA 
-- 
34.76±0.94c  NA 
-- 
NA 
-- 
NA 
-- 
NA 
-- 
10.000 fg  NA 
NA 
1.000 fg 
100 fg 
NA 
27.75±0.02a  28.15±0.05a  NA 
-- 
-- 
31.91±0.38b  32.82±0.47b  NA 
34.67±0.41c  38.65±0.15*  NA 
-- 

27.21±0.69a 
31.11±1.1b 
35.64±1.2c 

P. 
humuli 

P. 
cubensis 
clade II 
-- 
-- 
-- 
10.000 fg 
1.000 fg 
100 fg 
-- 
-- 
-- 

P. 
cubensis 
clade I  
10.000 fg 
1.000 fg 
100 fg 
-- 
-- 
-- 
-- 
-- 
-- 
10.000 fg  10.000 fg 
1.000 fg 
1.000 fg 
100 fg 
100 fg 
10.000 fg  10.000 fg  NA 
-- 
-- 
1.000 fg 
NA 
100 fg 
-- 
NA 
10.000 fg 
-- 
-- 
1.000 fg 
100 fg 
-- 

1.000 fg 
100 fg 
10.000 fg  27.39±0.07a  NA 
31±0.28* 
1.000 fg 
NA 
100 fg 
34.73±0.5c  NA 

28.58±0.15a  26.57±0.31a 
33.27±0.12*  31.25±0.05b 
NA** 
34.71±0.7c 
27.85±0.22a 
31.33±0.64b 
35.73±0.34c 

 

104 

 

Table 3-2. (cont’d) 
 

Species/Clade DNA 

Probe 

Mix 
ratio 

1/10 

1/100 

Pcub2 

P. 
humuli 

P. 
P. 
cubensis 
cubensis 
clade I  
clade II 
10.000 fg  1.000 fg 
1.000 fg 
10.000 fg 
10.000 fg  1.000 fg 
-- 
-- 
1.000 fg 
-- 
10.000 fg 

Pcub_ 
RFLP_qP1 
Cq±SD 
27.61±0.01a  31.9±0.56b  NA 
NA** 
28.21±0.18a  NA 
28.24±0.52a  29.51±0.13* 
NA 
10.000 fg  NA 
33.24±1.77b  26.32±0.32a 
31.06±1.11b 
NA 
1.000 fg 

Phum_ 
RFLP_qP4 
Cq±SD 

Cq±SD 

-- 
-- 

27.78±0.33 
a 

-- 

1.000 fg 
10.000 fg  100 fg 
100 fg 
10.000 fg 
100 fg 
-- 
-- 

10.000 fg 
-- 
-- 
10.000 fg  100 fg 
100 fg 

10.000 fg  34.25±0.48*  NA 
-- 
-- 
100 fg 
10.000 fg  NA** 

27.87±0.25a 
NA 
28.11±0.05a  NA 

27.66±0.06a  NA** 
NA** 
35.76±0.26c 
27.61±0.17a  NA 
NA 
27.92±0.55a 
28.08±0.78a  33.04±0.9* 
NA** 
26.65±0.52* 

NA 
10.000 fg  NA 

 
Cq values with the same letter are not significantly different. Cq values with the symbol * are significantly different from 
values obtained for un-mixed samples with the same concentration of DNA (t-Test; P=0.05). --: not DNA added. NC: not 
used for comparison in the t-Test. NA: not defined.

 

 

105 

 

Table 3-3. Percentage of qPCR positive samples collected by Burkard and 
impaction spore traps   

Location 

Year 

Burkard 
24 ha 
Reel  Nd 

Positivee 

Impaction 

96 hc 

Muskegon 

Bay 

Allegan 

Saginaw 

Ingham 

Average 

2018  15 
2019  16 
2018  15 
2018  16 
2018  15 
2019  16 
2018  15 
2019  16 
2018 
-- 
2019  16 

105  40.95% 
112  33.04% 
105  39.05% 
112  43.75% 
105  48.57% 
112  33.93% 
105  24.76% 
112  41.96% 
-- 
112  70.54% 
15.6  108  41.84% 

-- 

 
  24 hb 
  Nd 
  45 
  45 
  45 
  45 
  45 
  45 
  48 
  45 
-- 
 
  43 
  45 

Positive  Nd 
33.33% 
16 
15 
8.89% 
16 
13.33% 
20.00% 
15 
15 
26.67% 
15 
20.00% 
12.50% 
16 
15 
28.89% 
-- 
-- 
15 
46.51% 
23.35% 
15 

Positivee 
25.00% 
40.00% 
18.75% 
40.00% 
13.33% 
33.33% 
31.25% 
20.00% 
-- 
46.67% 
29.81% 

a. Reels were collected every 7 days and the tape was divided every 24 hours.   
b. Rod samples collected every 24 hours. 
c. Rod samples collected every 96 hours. 
d. Total number of samples collected. 
e. Percentage of samples collected that tested positive for P. cubensis or P. humuli. 

 

106 

 

 

Table 3-4. Linear regression analysis of Cq values as a function of the number Pseudoperonospora spp. 
sporangia 

Taxon 

P. humuli  

P. 
cubensis 
clade II 

b 
Variablesa  
b0  32.33 
Intercept 
b1 
-4.77 
log10(Xb) 
qPCR(A/B) 
-0.77 
b2 
log10(Xb)*qPCR (A/B) c  b3  0.71 
b0  33.69 
Intercept 
b1 
log10(Xb) 
-5.34 
qPCR (A/B) 
-2.30 
b2 
log10(Xb)*qPCR(A/B) c  b3  2.27 

Estimate  S.E 
0.42 
0.47 
0.55 
0.63 
0.59 
0.49 
0.65 
0.53 

t value 
76.13 
-10.05 
-1.40 
1.12 
57.03 
-10.89 
-3.50 
4.28 

Pr(>|t|) 
<2e-16 
<2e-16 
0.16 
0.26 
<2e-16 
<2e-16 
5.1e-04 
2.2e-05 

Pr(>F) 
-- 
<2e-16 
0.39 
0.26 
-- 
<2.20e-16 
0.91 
<2.2e-05 

a. Linear regression equation:  !"!.#$%= b&− b'	&'('&(*+',-.(/-)+ b(	&'('&("2!3)+ b)	&'('&(*+',-.(/-)("2!3	), 
!"!.*$= b&− b'	&'('&(*+',-.(/-)). 

b. X = number of sporangia. 
c. qPCR assay express as a binary variable: qPCR assay A = 1, qPCR assay B = 0. 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

 

107 

 

Table 3-5. b estimates of logistical models developed to predict the probability of Pseudoperonospora cubensis 
detection in Burkard and impaction spore trap samples 

Spore trap  b0 
Burkard 
Impaction 

-1.57 
-2.33 

SD 
0.10 
0.21 

Pvalue 
<2e-16 
<2e-17 

b1 
2.16 
1.60 

SD 
0.16 
0.21 

Pvalue 
<2e-16 
6.e-14 

AIC 
1333.8 
405.4 

Logistic regression models were developed using the spore trapping data collected in Allegan, Muskegon, Saginaw, 
Berrien and Bay counties in 2018 and 2019. Cq values of the qPCR protocol A were categorized as a binary variable 
using a threshold value of 35.5 (0>35.5 and 1≤35.5). To predict the probability of P. cubensis detection in Burkard and 
impaction spore traps as a function of the number of sporangia b estimates should be replaced in equation 1. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

 

108 

 

Table 3-6. Detection of Pseudoperonospora cubensis using Burkard spore traps coupled with qPCR and 
detection of symptoms  

Year  Location  qPCR   Spore trap 

CDM 

Symptoms  

qPCR detection of P. cubensis pre-

Burkard 

Aug 7 

Jul 23, 24, 26, 28, 29, 30, 31 and Aug 1, 

Jul 27, 28, 29, 30, 31and Aug 1, 2, 3, 4, 5, 

symptomsa   

2, 3, 4, 5, 6  

6  

Muskegon 

A  

2018 

Bay 

Allegan  

Saginaw 

Muskegon 

Bay 

2019 

Allegan  

Saginaw 

Ingham 

A  

A  

A  

B 

B 

B 

B 

B 

Impaction 
Burkard 
Impaction 
Burkard 
Impaction 
Burkard 
Impaction 
Burkard 
Impaction 
Burkard 
Impaction 
Burkard 
Impaction 
Burkard 
Impaction 
Burkard 
Impaction 

 

Aug 7 

 

Aug 21 

 

Aug 2 

 

Aug 16 

 

Aug 20 

 

Aug 16 

 

Aug 22 

 

Aug 21 

 

 Jul 27, 30 and Aug 5, 6 

--- 

Aug 4, 6, 7, 8, 9, 10, 11, 12, 13, 14  

Aug 2, 3, 4, 5, 6, 7, 14  

Jul 27, 31 

--- 

Aug 5, 6, 7, 9, 11, 12, 13, 14, 15  
Aug 8, 9, 10, 11, 12, 13, 14, 15  

Aug 12 

--- 

Aug 5, 8, 11, 12, 13, 14, 15  

Aug 8, 9, 10, 11, 15 
Aug 11, 19, 21   

Aug 15, 16, 17, 18, 19    

Aug 1, 19, 20 

--- 

qPCR detectionb  

16 days 

12 days  
11 days 

--- 

17 days 
19 days 
6 days 

--- 

11 days 
9 days  
8 days 

--- 

11 days 
9 days  
11 days 
9 days 
21 days 

--- 

a. qPCR detection within a window of three weeks before symptoms were observed in the field 
b. Number of days between the first detection of P. cubensis DNA using qPCR and the detection of CDM symptoms in the 
field  

 

 

109 

 

A                                                                             B

n
a
hig
e Mic

k
a
L

Muskegon Co.

Allegan Co.

L

a

k

e

 

H

u

r

o

n

n
a
hig
e Mic

k
a
L

Bay Co.

Berrien Co.

L

a

k

e

 

H

u

r

o

n

Saginaw Co.

Ingham Co.

 
Figure 3-1. Location of spore traps by county in 2018 (A) and 2019 (B). Green and blue dots indicate Burkard and 
impaction spore traps, respectively.  
 

 

 

 

 

110 

 

 

Figure 3-2. Standard curves for the quantification of Pseudoperonospora 
cubensis clade I, P. cubensis clade II and P. humuli DNA using the qPCR protocol 
A.  
Standard curves of the protocol A using (A) the probe CUBprobeSNP105 for P. 
cubensis detection and (B) the probe HUMprobeSNP105 for P. humuli detection.  The 
curves were constructed assessing the DNA from all taxa independently. Detection of P. 
cubensis clade I and II was observed using the probe HUMprobeSNP105. The log10 of 
DNA fg is plotted against the quantification cycle values (Cq). All data points are from 
three technical replicates derived from two DNA extractions. The centerline represents 
the line of fit and error bars represent standard deviation.  

 

111 

 

 

Figure 3-3. Standard curves for the quantification of Pseudoperonospora cubensis clade I, P. cubensis clade II 
and P. humuli DNA using the qPCR protocol B 
Standard curves of the protocol B using the probes Pcub2 (A), Pcub_RFLP_qP1 (B) and Phum_RFLP_qP4 (C) for 
detection of P. cubensis clade II, P. cubensis clade I and P. humuli, respectively.  The curves were constructed assessing 
the DNA from all taxa independently. Amplification of non-target taxa was not observed. The log10 of DNA fg is plotted 
against the quantification cycle values (Cq). All data points are from three technical replicates derived from two DNA 
extractions. The centerline represents the line of fit and error bars represent standard deviation.  

 

 

 

112 

 

A

q
C

35

30

25

20

0

B

Y = 31 – 2.9 log10(X)
Y = 33 - 5 log10(X)

q
C

qPCR A (SPcHx) R2=0.37
qPCR B (qPc2Hx) R2=0.60

1

3
Log10 (number of sporangia)

2

35

30

25

20

0.0

Y = 32 – 4.29 log10(X)
Y = 32 – 4.1 log10(X)

qPCR A (qPhRd ) R2=0.27
qPCR B (SPhFa ) R2=0.30

0.5
2.0
Log10 (number of sporangia)

1.0

1.5

2.5

Figure 3-4. Linear regression of Cq values as a function of sporangial numbers.  
(A)  Linear regression of Pseudoperonospora spp. sporangia counted using light microscopy against corresponding mean 
Cq values of the protocol A (rhomboids) and B (triangles). (B) Linear regression of Pseudoperonospora spp. sporangia 
counted using light microscopy against corresponding mean Cq values of the protocol A (circles) and B (triangles). The 
log10 number of sporangia is plotted against the quantification cycle values (Cq). 
 
 

 

 

 

113 

 

A

y
t
i
l
i

b
a
b
o
r
P

C

i

)
a
g
n
a
r
o
p
s

 
f
o
 
r
e
b
m
u
n
(
 

0
1
g
o
L

4

3

2

1

0

1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0

0
1
Negative/Positive

D

4

3

t
s
e
T

2

1

0

C

i

)
a
g
n
a
r
o
p
s

 
f
o
 
r
e
b
m
u
n
(
 

0
1
g
o
L

4

3

2

1

0

0
1
0
1
Negative/Positive
Negative/Positive

y
t
i
l
i

b
a
b
o
r
P

B

D

4

3

t
s
e
T

2

1

0

0
1
0
1
Negative/Positive
Negative/Positive

0
1
Negative/Positive

C
D

i

)
a
g
n
a
r
o
p
s
 
f
o
t
 
s
r
e
e
T
b
m
u
n
(
 
0
1
g
o
L

4
4

3
3

2
2

1
1

0
0

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0.0

0.0

2.0
Log10 (number of sporangia)

1.5

0.5

1.0

0.0

2.5
Log10 (number of sporangia)

2.0

1.5

0.5

1.0

 

Figure 3-5. Logistic regression of qPCR results from Burkard (A) and impaction (B) spore traps as a function of sporangial 
numbers. 

 

 

114 

 

Figure 3-6. Monitoring of Pseudoperonospora cubensis and P. humuli sporangia using 
spore traps in Muskegon (A) and Bay (B) counties in 2018 (Location B).  

 

 

115 

 

 

Figure 3-6. (cont’d) 
 
The data from each county was divided into three rows of panels. The first row panels 
represent the daily sporangial numbers estimated through the analysis of Burkard spore 
trap samples using light microscopy (blue bars). The y-axis was trimmed to 40 
sporangia to facilitate the visualization of low counts. The middle panels represent the 
qPCR results of protocol A for the detection of P. cubensis (red bars) and P. humuli 
(green circles) in the tape of Burkard spore traps. The bottom panels represent the 
qPCR results of protocol A for the detection of P. cubensis (red bars) and P. humuli 
(green circles) in the rods of impaction spore traps. A black arrow denotes the 
monitoring starting date. The dashed line denotes the date of cucurbit downy mildew 
symptom detection in the field. Scouting efforts to detect CDM symptoms in growing 
cucumber regions are increased once sporangial loads exceed 10 sporangia/day. Bars 
below the x-axis denote the time the reels and rods were changed from each trap. 

 

116 

 

 

Figure 3-7. Monitoring of Pseudoperonospora cubensis and P. humuli sporangia using 
spore traps in Allegan (A) and Saginaw (B) counties in 2018.  

  

 

117 

 

 

Figure 3-7. (cont’d) 
 
The data from each county was divided into three rows of panels. The first row panels 
represent the daily sporangial numbers estimated through the analysis of Burkard spore 
trap samples using light microscopy (blue bars). The y-axis was trimmed to 40 
sporangia to facilitate the visualization of low counts. The middle panels represent the 
qPCR results of protocol A for the detection of P. cubensis (red bars) and P. humuli 
(green circles) in the tape of Burkard spore traps. The bottom panels represent the 
qPCR results of protocol A for the detection of P. cubensis (red bars) and P. humuli 
(green circles) in the rods of impaction spore traps. A black arrow denotes the 
monitoring starting date. The dashed line denotes the date of cucurbit downy mildew 
symptoms detection in the field. Scouting efforts to detect CDM symptoms in growing 
cucumber regions are increased once sporangial loads exceed 10 sporangia/day. Bars 
below the x-axis denote the time the reels and rods were changed from each trap.

 

118 

 

 

Figure 3-8. Monitoring of Pseudoperonospora cubensis and P. humuli sporangia using 
spore traps in Muskegon (A) and Bay (B) counties in 2019. 

 

 

119 

 

 

Figure 3-8. (cont’d) 
 
The data from each county was divided into three rows of panels. The first row panels 
represent the daily sporangial numbers estimated through the analysis of Burkard spore 
trap samples using light microscopy (blue bars). The y-axis was trimmed to 40 
sporangia to facilitate the visualization of low counts. The middle panels represent the 
qPCR results of protocol B for the detection of P. cubensis (red bars) and P. humuli 
(green circles) in the tape of Burkard spore traps. The bottom panels represent the 
qPCR results of protocol B for the detection of P. cubensis (red bars) and P. humuli 
(green circles) in the rods of impaction spore traps. A black arrow denotes the 
monitoring starting date. The dashed line denotes the date of cucurbit downy mildew 
symptoms detection in the field. Scouting efforts to detect CDM symptoms in growing 
cucumber regions are increased once sporangial loads exceed 10 sporangia/day. Bars 
below the x-axis denote the time the reels and rods were changed from each trap. 
 

 

120 

 

 

Figure 3-9. Monitoring of Pseudoperonospora cubensis and P. humuli sporangia using 
spore traps in Allegan (A) and Saginaw (B) counties in 2019. 

 

 

121 

 

 

Figure 3-9. (cont’d) 
 
The data from each county was divided into three rows of panels. The first row panels 
represent the daily sporangial numbers estimated through the analysis of Burkard spore 
trap samples using light microscopy (blue bars). The y-axis was trimmed to 40 
sporangia to facilitate the visualization of low counts. The middle panels represent the 
qPCR results of protocol B for the detection of P. cubensis (red bars) and P. humuli 
(green circles) in the tape of Burkard spore traps. The bottom panels represent the 
qPCR results of protocol B for the detection of P. cubensis (red bars) and P. humuli 
(green circles) in the rods of impaction spore traps. A black arrow denotes the 
monitoring starting date. The dashed line denotes the date of cucurbit downy mildew 
symptoms detection in the field. Scouting efforts to detect CDM symptoms in growing 
cucumber regions are increased once sporangial loads exceed 10 sporangia/day. Bars 
below the x-axis denote the time the reels and rods were changed from each trap. 

 

122 

 

 

Figure 3-10. Monitoring of P cubensis and P. humuli sporangia using spore traps in 
Ingham (A) and Berrien (B) counties in 2019.  

 

 

123 

 

 

Figure 3-10. (cont’d) 
 
The data from Ingham county was divided into five rows of panels. The first row 
panels represent the daily sporangial numbers estimated through the analysis of 
Burkard spore trap samples using light microscopy (blue bars). The y-axis was trimmed 
to 40 sporangia to facilitate the visualization of low counts. The middle second and 
third row panels represent the qPCR results of protocol A (row 2) and B (row 3) for the 
detection of P. cubensis (red bars) and P. humuli (green circles) in the tape of Burkard 
spore traps. The fourth and five row panels represent the qPCR results of protocol A 
(row 4) and B (row 5) for the detection of P. cubensis (red bars) and P. humuli (green 
circles) in the rods of impaction spore traps. The data from Berrien county was divided 
into three row panels. The first row panels represent the daily sporangial numbers 
estimated through the analysis of Burkard spore trap samples using light microscopy 
(blue bars). The y-axis was trimmed to 40 sporangia to facilitate the visualization of low 
counts. The middle panels represent the qPCR results of protocol A for the detection 
of P. cubensis (red bars) and P. humuli (green circles) in the tape of Burkard spore 
traps. The bottom panels represent the qPCR results of protocol B for the detection of 
P. cubensis (red bars) and P. humuli (green circles) in the tape of Burkard spore traps. 
A black arrow denotes the monitoring starting date. The dashed line denotes the date of 
cucurbit downy mildew symptoms detection in the field. Scouting efforts to detect CDM 
symptoms in growing cucumber regions are increased once sporangial loads exceed 10 
sporangia/day. Bars below the x-axis denote the time the reels and rods were changed 
from each trap. 

 

124 

 

 

 

 
Figure 3-11. Standard curves for the quantification of Pseudoperonospora 
cubensis and P. humuli sporangia using the qPCR protocol A.  
Five-point standard curves for protocol A using (A) the probe CUBprobeSNP105 
for P. cubensis detection and (B) the probe HUMprobeSNP105 for P. humuli 
detection. All data points represent an average of 8 to 15 replicate spore-count 
samples of each species. Error bars on points represent the standard error of the 
mean. There was inconsistent amplification for the 1-count sporangium samples 
and they were excluded from the final linear regression. 

 

125 

 

 

 
Figure 3-12. Standard curves for the quantification of Pseudoperonospora cubensis clade I, P. cubensis clade II 
and P. humuli sporangia using the qPCR protocol B.  
Five-point standard curves for protocol B using the probes Pcub2 (A), Pcub_RFLP_qP1 (B) and Phum_RFLP_qP4 (C) for 
detection of P. cubensis clade II, P. cubensis clade I and P. humuli, respectively. All data points represent an average of 8 
to 15 replicate spore-count samples of each taxa. Error bars on points represent the standard error of the mean. There 
was inconsistent amplification for the 1-count sporangium samples and they were excluded from the final linear 
regression. 

 

 

 

126 

 

 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

 

 

 

 

 

 

 

 

LITERATURE CITED 

127 

 

LITERATURE CITED 

 
 
Adams, M., D’Arcangelo, K, Collins, H, Quesada-Ocampo, L (2019) Evaluation of 
cultivars for control of downy mildew on squash, Kinston 2018. Plant Dis Manag 
Reports 13: V124 
 
Aylor, D.E. (2017) Chapter 20: Aerial Sampling. Aerial dispersal of pollen and spores. 
The American Phytopathological Society, pp 339–360 
 
Aylor, D.E. (1993) Relative collection efficiency of Rotorod and Burkard spore samplers 
for airborne Venturia inaequalis ascospores. Phythopathology 83: 1116–1119 
 
Bello, J.C., Sakalidis, M.L., Perla, D.E., Hausbeck, M.K. (2020) Detection of airborne 
sporangia of Pseudoperonospora cubensis in Michigan using Burkard spore traps 
coupled to qPCR. Accepted in Plant Dis.  
 
Blum, M., Waldner, M., Olaya, G., Cohen, Y., Gisi, U., Sierotzki, H. (2011) Resistance 
mechanism to carboxylic acid amide fungicides in the cucurbit downy mildew 
pathogen Pseudoperonospora cubensis. Pest Manag Sci 67: 1211–1214 
 
Carisse, O., McNealis, V., Kriss, A. (2017) Association between weather variables, 
airborne inoculum concentration, and raspberry fruit rot caused by Botrytis cinerea. 
Phytopathology 108: 70–82 
 
Carisse, O., Tremblay, D.M., Lévesque, C.A., Gindro, K., Ward, P., Houde, A. (2009) 
Development of a Taqman real-time PCR assay for quantification of airborne conidia 
of Botrytis squamosa and management of Botrytis leaf blight of onion. Phytopathology 
99: 1273–1280 
 
Choudhury, R.A., Koike, S.T., Fox, A.D., Anchieta, A., Subbarao, K.V., Klosterman, 
S.J., McRoberts, N. (2016a) Season-long dynamics of spinach downy mildew 
determined by spore trapping and disease incidence. Phytopathology 106: 1311–1318 
 
Choudhury, R.A., Koike, S.T., Fox, A.D., Anchieta, A., Subbarao, K.V., Klosterman, 
S.J., McRoberts, N. (2016b) Spatiotemporal patterns in the airborne dispersal of 
spinach downy mildew. Phytopathology 107: 50–58 
 
Coley-Smith, J.R. (1962) Overwintering of hop downy mildew Pseudoperonospora 
humuli (Miy. and Tak.) Wilson. Ann Appl Biol 50: 235–243 
 
Crandall, S. G., Ramon, M. L., Burkhardt, A. K., Bello J. C., Adair, N., Gent, D. H., 
Hausbeck, M. K., Quesada-Ocampo, L. M. and Martin, F. N. A multiplex TaqMan qPCR 
assay for detection and quantification of clade 1 and clade 2 isolates 
of Pseudoperonospora cubensis and P. humuli. Under review in Plant Dis. 
 

 

128 

 

Dhar, N., Mamo, B.E., Subbarao, K., Koike, S., Fox, A., Anchieta, A., Klosterman, S.J 
.(2019) Measurements of aerial spore load by qPCR facilitates lettuce downy mildew 
risk advisement. Plant Dis 104: 82–93 
 
Dung, J.K.S., Scott, J.C., Cheng, Q., Alderman, S.C., Kaur, N., Walenta, D.L., Frost, 
K.E., Hamm, P.B. (2018) Detection and quantification of airborne Claviceps 
purpurea sensu lato ascospores from Hirst-type spore traps using real-time quantitative 
PCR. Plant Dis 102: 2487–2493 
 
Fall. M.L., Van der Heyden, H., Brodeur, L., Leclerc, Y., Moreau, G., Carisse, O. (2015) 
Spatiotemporal variation in airborne sporangia of Phytophthora infestans: 
characterization and initiatives towards improving potato late blight risk estimation. Plant 
Pathol 64: 178–190 
 
Frenz, D.A. (1999) Comparing pollen and spore counts collected with the Rotorod 
Sampler and Burkard spore trap. Ann Allergy, Asthma Immunol 83: 341–349 
 
González-Domínguez, E., Berlanas, C., Gramaje, D., Armengol, J., Rossi, V., Berbegal, 
M. (2020) Temporal dispersal patterns of Phaeomoniella chlamydospora, causal agent 
of petri disease and esca, in vineyards. Phytopathology 110: 1216–1225 
 
Granke, L.L., Hausbeck, M.K. (2011) Dynamics of Pseudoperonospora 
cubensis sporangia in commercial cucurbit fields in Michigan. Plant Dis 95: 1392–1400 
 
Granke, L.L., Morrice, J.J., Hausbeck, M.K. (2013) Relationships between 
airborne Pseudoperonospora cubensis sporangia, environmental conditions, and 
cucumber downy mildew severity. Plant Dis 98: 674–681 
 
Hausbeck, M., Brisco-McCann, E., Goldenhar, K., Cook, A. (2017) Evaluation of 
fungicide programs for control of downy mildew of cucumber, 2016. Plant Dis Manag 
Reports 11: V063 
 
Hausbeck, M.K., Perla, D.E., Cook, A.J. (2019) Evaluation of single fungicide products 
for control of downy mildew of cucumber, 2018. Plant Dis Manag Reports 139: V13 
 
Hollomon, B. (2007) Fungicide resistance in crop pathogens: How can it be managed?, 
2nd ed. FRAC, Brussels 
 
Holmes, G.J., Ojiambo, P.S., Hausbeck, M.K., Quesada-Ocampo, L., Keinath, A.P. 
(2014) Resurgence of cucurbit downy mildew in the united states: a watershed event for 
research and extension. Plant Dis 99: 428–441 
 
Jackson, S.L., Bayliss, K.L. (2011) Spore traps need improvement to fulfil plant 
biosecurity requirements. Plant Pathol 60: 801–810 
 

 

129 

 

Keinath, A.P. (2015) Utility of a cucumber plant bioassay to assess fungicide efficacy 
against Pseudoperonospora cubensis. Plant Dis 100: 490–499 
 
Klosterman, S.J., Anchieta, A., McRoberts, N., Koike, S.T., Subbarao, K.V.,  
Voglmayr, H., Choi, Y-J., Thines, M., Martin, F.N. (2014) Coupling spore traps and 
quantitative PCR assays for detection of the downy mildew pathogens of spinach 
(Peronospora effusa) and beet (P. schachtii). Phytopathology 104: 1349–1359 
 
Kunjeti, S.G., Anchieta, A., Martin, F.N., Choi, Y-J., Thines, M., Michelmore, R.W., 
Koike, S.T., Tsuchida, C., Mahaffee, W., Subbarao, K. V., et al (2016) Detection and 
quantification of Bremia lactucae by spore trapping and quantitative PCR. 
Phytopathology 106: 1426–1437 
 
Lizotte, E., and Miles, T.D. (2020) Hop disease management in Michigan for 2020. MSU 
Ext., https://www.canr.msu.edu/news/managing-hop-downy-mildew-in-michigan 
 
Naegele, R.P., Quesada-Ocampo, L.M., Kurjan, J.D., Saude, C., Hausbeck, M.K. 
(2016) Regional and temporal population structure of Pseudoperonospora cubensis in 
Michigan and Ontario. Phytopathology 106: 372–379 
 
O’Hara, R., Tedone, E., Ludlow, A., Huang, E., Arosio, B., Mari, D., Shay, J.W. (2019) 
Quantitative mitochondrial DNA copy number determination using droplet digital PCR 
with single-cell resolution. Genome Res 29: 1878–1888 
 
Ojiambo, P.S., Gent, D.H., Quesada-Ocampo, L.M., Hausbeck, M.K., Holmes, G.J. 
(2015) Epidemiology and population biology of Pseudoperonospora cubensis: a model 
system for management of downy mildews. Annu Rev Phytopathol 53: 223–246 
 
Ojiambo, P.S., Holmes, G.J. (2010) Spatiotemporal spread of cucurbit downy mildew in 
the Eastern United States. Phytopathology 101: 451–461 
 
Ojiambo, P.S., Holmes, G.J., Britton, W., Babadoost, M., Bost, S.C., Boyles, R., Brooks, 
M., Damicone, J., Draper, M.A., Egel, D.S., et al (2011) Cucurbit downy mildew 
IPMpipe: a next generation web-based interactive tool for disease management and 
extension outreach. Plant Heal Prog 12: 26 
 
Perla, D., Hausbeck, M.K., Cook, A. (2019) Evaluation of fungicide programs for control 
of downy mildew of cucumber, 2018. Plant Dis Manag Reports 147: V13 
 
Rahman, A., Standish, J., D’Arcangelo, K.N., Quesada-Ocampo, L.M. (2020) Clade-
specific biosurveillance of Pseudoperonospora cubensis using spore traps for precision 
disease management of cucurbit downy mildew. Phytopathology. 
 
Reuveni, M., Eyal, H., Cohen, Y. (1980) Development of resistance to metalaxyl 
in Pseudoperonospora cubensis. Plant Dis 64: 1108–1109 
 

 

130 

 

Rogers, S.L., Atkins, S.D., West, J.S. (2009) Detection and quantification of airborne 
inoculum of Sclerotinia sclerotiorum using quantitative PCR. Plant Pathol 58: 324–331 
 
Savory, E.A., Granke, L.L., Quesada-Ocampo, L.M., Varbanova, M., Hausbeck, M.K., 
Day, B. (2011) The cucurbit downy mildew pathogen Pseudoperonospora cubensis. Mol 
Plant Pathol 12: 217–26 
 
Solomon, W.R., Burge, H.A., Boise, J.R., Becker, M. (1980) Comparative particle 
recoveries by the retracting rotorod, rotoslide and Burkard spore trap sampling in a 
compact array. Int J Biometeorol 24: 107–116 
 
Summers, C., Adair, N., Gent, D., Mcgrath, M., Christine, S. (2015a) 
Pseudoperonospora cubensis and P. humuli detection using species-specific probes 
and high definition melt curve analysis. Can J Plant Pathol 37: 150605144651006 
 
Summers, C.F., Gulliford, C.M., Carlson, C.H., Lillis, J.A., Carlson, M.O., Cadle-
Davidson, L., Gent, D.H., Smart, C.D. (2015b) Identification of genetic variation between 
obligate plant pathogens Pseudoperonospora cubensis and P. humuli using RNA 
sequencing and genotyping-by-sequencing. PLoS One 10: e0143665 
 
Thiessen, L.D., Neill, T.M., Mahaffee, W.F. (2017) Timing fungicide application intervals 
based on airborne erysiphe necator concentrations. Plant Dis 101: 1246–1252 
 
Thomas, A., Carbone, I., Choe, K., Quesada-Ocampo, L.M., Ojiambo, P.S. (2017) 
Resurgence of cucurbit downy mildew in the United States: Insights from comparative 
genomic analysis of Pseudoperonospora cubensis. Ecol Evol 7: 6231–6246 
 
USDA (2020) United States Department of Agriculture. 2019. Vegetables Annual 
Summary.  
 
Van der Heyden, H., Bilodeau, G., Carisse, O., Charron, J-B. (2020) Monitoring 
of Peronospora destructor primary and secondary inoculum by Real-Time qPCR. Plant 
Dis. doi: 10.1094/PDIS-03-20-0687-RE 
 
Villari, C., Mahaffee, W.F., Mitchell, T.K., Pedley, K.F., Pieck, M.L., Hand, F.P. (2016) 
Early detection of airborne inoculum of Magnaporthe oryzae in turfgrass fields using a 
quantitative LAMP assay. Plant Dis 101: 170–177 
 
Wallace, E.C., D’Arcangelo, K.N., Quesada-Ocampo, L.M. (2020) Population analyses 
reveal two host-adapted clades of Pseudoperonospora cubensis, the causal agent of 
cucurbit downy mildew, on commercial and wild cucurbits. Phytopathology. doi: 
10.1094/PHYTO-01-20-0009-R 
 
West, J., Kimber, R. (2015) Innovations in air sampling to detect plant pathogens. Ann 
Appl Biol 166: 4–17 
 

 

131 

 

Wyka, S.A., McIntire, C.D., Smith, C., Munck, I.A., Rock, B.N., Asbjornsen, H., Broders, 
K.D. (2017). Effect of climatic variables on abundance and dispersal of Lecanosticta 
acicola spores and their impact on defoliation on eastern white pine. Phytopathology 
108: 374–383 
 
Zhu, S.S., Liu, X.L., Wang, Y., Wu, X.H., Liu, P.F., Li, J.Q., Yuan, S.K., Si, N.G. (2007) 
Resistance of Pseudoperonospora cubensis to flumorph on cucumber in plastic houses. 
Plant Pathol 56: 967–975 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

132 

 

 

 

 

 

 

 

 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

CHAPTER IV: GENOTYPING OF THE OBLIGATE PLANT PATHOGENS 
PSEUDOPERONOSPORA CUBENSIS AND P. HUMULI USING TARGET 

ENRICHMENT SEQUENCING 

133 

 

ABSTRACT  

Technological advances in genome sequencing have improved our ability to 

catalog genomic variation and have led to an expansion of the scope and scale of 

genetic studies over the past decade. Yet, for agronomically important plant pathogens 

such as the downy mildews (Peronosporaceae) the scale of genetic studies remains 

limited. This is, in part, due to the difficulties associated with maintaining obligate 

pathogens, and the logistical constraints involved in the genotyping of these species 

(e.g. obtaining DNA of sufficient quantity and quality). To gain an evolutionary and 

ecological perspective of downy mildews, adaptable methods for the genotyping of their 

populations are required. Here, we describe a targeted enrichment (TE) protocol to 

genotype isolates from two Pseudoperonospora species (P. cubensis and P. humuli) 

using less than 50 ng of mixed pathogen and plant DNA for library preparation. We 

were able to enrich 736 target genes across 101 samples and identified 2,978 high-

quality SNP variants. Using these SNPs, we detected significant genetic differentiation 

(AMOVA, p=0.01) between P. cubensis subpopulations from Cucurbita moschata (clade 

I) and Cucumis sativus (clade II) in Michigan (U.S.). No evidence of location-based 

differentiation was detected within the P. cubensis (clade II) subpopulation of Michigan. 

However, a significant effect of location on the genetic variation of the P. humuli 

subpopulation was detected in the state (AMOVA, p=0.01). Mantel tests found evidence 

that the genetic distance among P.  humuli samples was associated with the physical 

distance of the hop yards from which the samples were collected (p=0.005). The 

differences in the distribution of genetic variation of the P. humuli and P. cubensis 

subpopulations of Michigan suggest differences in the dispersal of these two species. 

 

134 

 

The TE protocol described here provides an additional tool for genotyping obligate 

biotrophic plant pathogens and the execution of new genetic studies. 

INTRODUCTION  

Downy mildew (DM) pathogens (Peronosporaceae) cause foliar disease in 

several agronomically important plant species (Gent et al., 2009; Kanetis et al, 2013; 

Kunjeti et al., 2016; Lane et al., 2005; Rivera et al., 2016; Wallace & Quesada-Ocampo, 

2017; Wong & Wilcox, 2001). The group is comprised of at least twenty different genera 

representing a large portion of the described pathogenic species of oomycetes (Thines, 

2014; Choi and Thines, 2015). All DM species are considered obligate biotrophs, 

growing only in association with living host tissue (Bourret et al., 2018). Over the past 

decade, several studies have provided key insight into DM biology, including virulence 

mechanisms (Savory et al., 2012; Baxter et al., 2010; Burkhardt & Day, 2016; 

Purayannur et al., 2020), host specificity (Choi and Thines, 2015; Summers et al., 

2015b; Rivera et al., 2016; Wallace et al., 2020), and fungicide resistance (Gisi and 

Sierotzki, 2008; Blum et al., 2011). Despite this progress, several unresolved research 

questions in ecology and evolution remain, many of which could be addressed with 

emerging genomic and genetic approaches.  

Recent technological advances in high-throughput sequencing have facilitated 

the genotyping of dozens of samples and the analysis of whole plant-pathogen 

genomes (Savory et al., 2012b; Sharma et al., 2015; Rivera et al., 2016; Withers et al., 

2016; Cui et al., 2019; Fletcher et al., 2019; Rahman et al., 2019). The recent 

availability of genome-wide sequence information from multiple individuals of the same 

species has facilitated population genomic studies of several plant pathogens 

 

135 

 

(Grünwald et al., 2017; Tabima et al., 2018; Carleson et al., 2019; Gent et al., 2019). 

Whole or reduced representation genome sequencing technologies have become 

increasingly popular to monitor genetic changes in plant-pathogen populations. 

However, these technologies have not been broadly used to study DM populations due 

to difficulties in obtaining sufficient amounts of DNA from single obligate biotrophic 

individuals (Milgroom, 2015).  

Traditionally, other genotyping technologies that require significantly lower inputs 

of DNA such as Sanger sequencing and microsatellites (SSRs) are used with DNA 

extracted from fresh and stored symptomatic tissue to study DM populations (Quesada-

Ocampo et al., 2012; Kitner et al., 2015; Naegele et al., 2016; Rivera et al., 2016; 

Wallace and Quesada-Ocampo, 2017; Wallace et al., 2020). However, these 

approaches are limited by the relatively low number of variants that can be obtained 

when compared to genome-wide sequencing options. DNA from symptomatic tissue 

can be used for sequencing with high-throughput sequencing technologies but the low 

concentration of pathogen DNA and large amounts of exogenous material (e.g. plant 

and bacterial DNA) can add further expense and complexity to the sequencing and 

bioinformatic analysis (Stassen et al., 2012).  

The quality of high-throughput sequencing data depends on the quality and purity 

of the DNA sample, which means that the target pathogen is ideally present in higher 

amounts when compared to the host plant before DNA extraction (Jouet et al., 2019). A 

spore propagation approach using detached susceptible leaves can be used to 

propagate bulk amounts of DM sporangia from which additional pathogen DNA can be 

recovered (Summers et al., 2015b; Thomas et al., 2017a; Gent et al., 2019); however, 

 

136 

 

multiple growing cycles (7-10 days/cycle) are needed making this approach time 

consuming and labor-intensive (Ali et al., 2011; Gent et al., 2019). Spore propagation is 

most successful when using freshly collected symptomatic leaves, but propagating 

sporangia from samples stored over long periods of time or under poor conditions (e.g. 

samples stored >-80 C) can be complicated. Additionally, sporangial propagation 

requires continuous maintenance of fresh sporangia on a highly susceptible host, which 

may bias the genetic composition of the populations under study due to the selection of 

genotypes by the host and the propagation conditions (Jones et al., 2014; Thomas et 

al., 2017a). While sporangial propagation can provide large amounts of high-quality 

DNA needed for sequencing, this approach is impractical for population studies 

requiring a large number of samples collected over time (i.e. years).  

 

Sequence capture methods may provide a solution for genotyping DM pathogens 

that excludes non-target DNA while also facilitating high coverage sequencing of 

several target loci (Kozarewa et al., 2015; Lim & Braun, 2016). These techniques use 

affinity probes (RNA/DNA) to isolate particular sequences of interest (“target regions”) 

out of a larger pool of DNA fragments (DNA library). Sequence capture methods such 

as target enrichment (TE) have been used across a variety of genomic studies with 

model (Gnirke et al., 2009; Clark et al., 2011) and non-model organisms (Faircloth et al., 

2015; McCormack et al., 2016; Starrett et al., 2017) and have also facilitated the study 

of museum specimens with low amounts of poor quality DNA (Cruz-Dávalos et al., 

2017). Additionally, sequence capture methods have also been used to study the 

genetic variation of other plant pathogens such as Phytophthora infestans, 

Phytophthora (cid:0)apsica (Thilliez et al., 2019), and the obligate biotroph Albugo candida 

 

137 

 

(Jouet et al., 2019). These techniques could facilitate the study of DM pathogen 

populations using symptomatic tissue samples, alleviating the need for sporangial 

propagation and expanding the type and condition of the samples used.  

In this study, we evaluated the genotypic variation of two DM species of the 

genus Pseudoperonospora using TE sequencing. Pseudoperonospora cubensis and 

Pseudoperonospora humuli infect cucurbits and hops, respectively, worldwide and are 

considered the most economically important species of the genus Pseudoperonospora 

(Choi et al., 2005; Mitchell et al., 2011). P. cubensis (clades I and II) causes foliar blight 

(DM) of cucumber (Cucumis sativus), melon (Cucumis melo), pumpkin (Cucurbita 

maxima), watermelon (Citrullus lanatus), and squash (Cucurbita moschata) (Summers 

et al., 2015b; Thomas et al., 2017a; Wallace et al., 2020). P. humuli negatively impacts 

hop (Humulus lupulus) cone yield (Gent et al., 2010). We describe an optimized TE 

procedure for sequencing DM and a bioinformatic pipeline for population genetic 

analyses using TE data. Evaluation of the structure, diversity, and reproduction of P. 

cubensis and P. humuli populations in Michigan was of particular interest. 

MATERIALS AND METHODS  

Sample collection and DNA extraction. Over a 12 year period (2007-2009, 

2012-2013, 2015, and 2018-2019). P. cubensis sporangia were obtained from 

symptomatic cucurbit tissue collected  in Michigan, other U.S. states (Indiana, Iowa, 

Ohio, Florida, and Wisconsin) and one Canadian province: Ontario (Fig. 4-1 A; 

Supplementary Table 4-1). Sporangia were harvested by gently rinsing infected leaves 

exhibiting multiple lesions and pathogen signs using a Preval spray power unit (Preval, 

Chicago) filled with distilled water as described by Mitchell et al. (2011). Dislodged 

 

138 

 

sporangia were transferred into a centrifuge tube (2 mL), pelleted (14000 rpm for 5 min, 

5424 Centrifuge, Eppendorf) and subjected to DNA extraction. A second group of 

samples was processed by excising single lesions using a sterile scalpel. Sporangia 

from each lesion were harvested by vigorously shaking the microcentrifuge tube (2 mL) 

containing the tissue and 1 mL of distilled H2O for 30 s. Dislodged sporangia were 

pelleted (14000 rpm for 5 min, 5424 Centrifuge, Eppendorf) and subjected to DNA 

extraction.  

P. humuli samples were processed using methods similar to those described by 

Chee et al. (2006). Sporangial suspensions were obtained from diseased basal shoots 

collected from six commercial hop yards and a hop research plot located in Michigan. 

The three hop yards in the north region were located within a radius of approximately 25 

km (hop yards A, B and C) (Fig. 4-1 B). The two hop yards in the west region of the 

state were separated by approximately 100 km and operated by the same producer 

(hop yards D and E) (Fig. 4-1 B). The sixth commercial hop yard was located in the east 

region (hop yard F) and was located approximately 80 km from the research plot (hop 

yard G) located in the central region (Fig. 4-1B).  Diseased basal shoots were brought 

to the laboratory, stems placed into beakers of water, and shoots individually covered 

with a plastic bag overnight to induce sporulation. The following day, sporangia were 

washed from the abaxial leaf surface using a Preval spray power unit filled with distilled 

water. Sporangial suspensions from single shoots were transferred into centrifuge tubes 

(2 ml) and pelleted by centrifugation (14000 rpm for 5 min). All DNA extractions were 

performed on the pelleted sporangial suspensions using the NucleoSpin Plant II 

 

139 

 

isolation kit (Macherey-Nagel, Bethlehem, PA, U.S.) following manufacturer’s 

instructions. 

DNA library construction.  

TE libraries: Whole DNA was extracted from 275 samples as described above and 

quantified using a Qubit 2.0 fluorimeter (Thermo Fisher Scientific). Before fragmentation 

of DNA using a M220 Focused-ultrasonicator (Covaris, Woburn, MA, U.S.) the presence 

of P. cubensis and/or P. humuli in the samples was confirmed by sanger sequencing of 

the ITS region as described by Quesada-Ocampo et al. (2012). A subset of five 

samples were divided in two individual tubes and submitted independently for library 

preparation and enrichment. These samples were used to estimate the number of errors 

introduced into a sample during the TE workflow. Sequencing libraries were prepared 

using the KAPA HyperPrep library sequencing preparation kit (Kapa/Roche, 

Pleasanton, CA, U.S.) following the manufacturer’s recommendations and an additional 

KAPA Pure Bead (Kapa/Roche, Pleasanton, CA, U.S.) cleanup to ensure the libraries 

were free from any adapters. Insert size and library quality was verified using the Agilent 

2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, U.S.). To optimize the use of 

the Mybaits custom kit’s reactions (80bp baits, Arbor Biosciences, formerly Mycroarray 

Inc., Ann Arbor, MI, U.S.), good quality libraries were pooled into groups of four libraries 

for enrichment. Probes (baits) were designed according to manufacturer protocols 

(probe compatibility, repeat masking, and melting temperature filters) to cover 118 

population informative DNA regions and 712 DNA regions annotated as protein genes 

at a 2X coverage. Protein gene sequences were obtained using bedtools (Quinlan and 

Hall, 2010) with the coordinates information derived from the genome annotation of P. 

 

140 

 

cubensis available online in the server of the Oregon State University Libraries 

(dx.doi.org/10.7267/N9TD9V7M; Burkhardt et al., 2014). All the genes annotated as 

extracellular toxins, hydrolytic enzymes, enzyme inhibitors, cell-entering RxLR effectors, 

cell-entering crinkler effectors, secreted proteins and transcription factors were targeted 

for sequencing. These regions were selected to identify annotated genes containing 

SNPs that may play a role in controlling host-specificity, and selectively neutral regions 

for population level analyses (Quesada-Ocampo et al., 2012; Kitner et al., 2015; 

Wallace and Quesada-Ocampo, 2017).   

Hybridization of DNA libraries and probes were performed at 65 °C for 24 or 48 h 

(Supplementary Table 4-1). After hybridization, library pools were bound to Dynabeads 

MyOne Streptavidin C1 magnetic beads (Life Technologies) for enrichment. 

Amplification of the enriched libraries was performed using the 2X KAPA HiFi HotStart 

ReadyMix (Kapa/Roche, Pleasanton, CA, U.S.) and 5 μM of each TruSeq forward and 

reverse primers. Amplification conditions were 98 °C for 2 min, followed by 6-14 cycles 

of 98 °C for 20 s, 60 °C for 30 s and 72 °C for 60 s and then a final extension of 72 °C 

for 5 min. Following PCR, libraries were quantified using a Qubit 2.0 fluorimeter 

(Invitrogen). Enriched libraries were quality controlled and quantified using a 

combination of Qubit dsDNA HS and Caliper LabChipGX HS DNA assays. Based on 

these quantifications the libraries were combined in equimolar amounts into two groups 

for sequencing. The first set of samples (TES1) included 62 enriched libraries with a 

hybridization time of 24 h. The second set (TES2) of samples included 72 enriched 

libraries with a hybridization time of 48 h (Supplementary Table 4-1). Both sets were 

quantified using the Qubit 2.0 and the Kapa Biosystems Illumina Library Quantification 

 

141 

 

qPCR kit. Each set was loaded onto one line of an Illumina HiSeq 4000 flow cell and 

sequenced in a 2x150bp paired-end format. 

Low-coverage whole-genome sequencing (Lc-WGS) libraries: In order to 

assess the performance of TE in comparison to Lc-WGS and to estimate the amount of 

exogenous material in the DNA samples, ten samples of genomic DNA directly 

extracted from sporangial suspensions derived from individual lesions were submitted to 

the Michigan State University Research Technology Support Facility (MSU RTSF) for 

DNA library preparation and sequencing. Due to the low DNA concentration of some of 

the samples, the Rubicon ThruPLEX DNA-Seq library preparation kit (Takara Bio, 

Mountain View, CA, U.S.) was used to prepare the sequencing libraries. Library 

preparation was performed following manufacturer’s recommendations with an 

additional AMPureXP bead cleanup (Beckman Coulter, Indianapolis, IN, U.S.). 

Completed libraries were quality controlled and quantified using a combination of Qubit 

dsDNA HS (Thermo Fisher Scientific/Life Technologies, Carlsbad, CA, U.S.) and 

Caliper LabChipGX HS DNA assays (PerkinElmer, Waltham, MA, U.S.). Based on 

these quantifications the libraries were grouped in equimolar amounts into two pools 

that were quantified using the Kapa Biosystems Illumina Library Quantification qPCR kit 

(Kapa/Roche, Pleasanton, CA, U.S.). Each pool was loaded into one lane of an Illumina 

MiSeq standard flow cell (v2) (Illumina, San Diego, CA, U.S.) and sequencing was 

performed in a 2x250bp paired end format using a v2 500 cycle MiSeq reagent cartridge 

(Illumina, San Diego, CA, U.S.). 

Genotyping and variant calling. In order to assess the performance of TE for 

the genotyping of P. cubensis and P. humuli, we analyzed the data in combination with 

 

142 

 

previously published whole-genome sequencing (WGS) data generated for a 

comparative genomic analyses of P. cubensis. This data set included a total of 10 

samples collected from different cucurbits in the U.S. states of Alabama, South Carolina 

(N=2), Florida, North Carolina (N=2), Georgia, California, New York and Oregon. The 

datasets can be found in the bioprojects PRJNA360426 (Thomas et al., 2017) and 

PRJNA675756 (this project) available online through the National Center for 

Biotechnology Information (NCBI). Sequencing raw reads of all libraries were trimmed 

and quality filtered using Trimmomatic 0.32 (Bolger et al., 2014). The publicly available 

P. cubensis (Savory et al., 2012b) and C. sativus genome sequences were used as 

references for alignment of the reads. To estimate plant contamination, high-quality 

reads were initially aligned to the C. sativus genome using the Burrows-Wheeler aligner 

(BWA-0.7.17) (Li and Durbin, 2009). All reads were then aligned to a reference genome 

of P. cubensis containing only the contigs where the target genes are located. Clean 

reads of the Bioproject PRJNA360426 were also aligned to this subset of the genome. 

Samtools-0.1.19  (Li et al., 2009) was used to sort the alignments and PCR duplicates 

were removed with Picard (https://broadinstitute.github.io/picard/). Coverage depth was 

calculated using BEDTools-2.27.1.  

The Next Generation Sequencing Plugin (NGSEP) for analysis of high-

throughput sequencing data was used for calling SNPs and producing the variant call 

format file (VCF) (Perea et al., 2016). The VCF file generated was filtered using the 

NGSEP plugin in order to retain the most informative SNPs of sufficient quality. Only 

biallelic SNPs, present in 70% of samples, with a minimum quality of 40 were retained. 

The resulting VCF file was imported into R using the package vcfR (version 1.5.0) 

 

143 

 

(Knaus and Grünwald, 2017). Missing genotype data were deleted before further 

analysis (Tabima et al., 2018). Samples with more than 40% of missing variants were 

removed from the analysis to retain the maximum number of SNPs. Variants with more 

than 75% of missing information across samples were also removed. To increase the 

stringency for the genotyping, we excluded all the samples with an average coverage 

below 50X. A nonparametric Kruskal-Wallis rank-sum test was used to assess the 

differences among the mean coverage depth across libraries.  

Population Assignment.  P. cubensis samples were assigned to several 

different subpopulations depending on the analyses. To study the distribution of the 

genetic variation by host, P. cubensis samples were assigned to subpopulations 

according to its host species. To study the distribution of the genetic variation by 

geographic location, only the P. cubensis samples collected from Cucumis spp. were 

used. The samples collected in the U.S. states of Michigan, Wisconsin, Indiana, Ohio, 

and the Canadian province of Ontario, CA, were assigned to the Midwest subpopulation 

(Fig, 4-1 C). Midwest samples excluding those from Michigan were assigned to the 

Midwest-w/o-MI subpopulation (Fig. 4-1 C). Samples from Michigan were further divided 

into three subpopulations according to the region of collection (east, west and central) 

(Fig. 4-1 C). Samples collected from Iowa, Alabama, Oregon, California, South 

Carolina, Florida, New York, North Carolina, and Georgia, were assigned to their 

respective state of origin.  

P. humuli samples were assigned to either the north or west subpopulation 

based on their location of collection in Michigan (Fig. 4-1 B, C; Table 4-2; 

Supplementary Table 4-1). Three hop yards (A, B and C) were located within a radius of 

 

144 

 

approximately 25 km and comprised the north subpopulation (Fig. 4-1 C). Hop yards D 

and E in southwest Michigan are operated by one producer and make up the west 

subpopulation (Fig. 4-1 C). Representative samples from central, east Michigan and 

Oregon were also included in our analysis.  

Genetic differentiation and population structure. The filtered SNPs were used 

to construct a genetic tree based on Nei’s genetic distance (Nei, 1972) and visualize the 

genetic relationship among all samples, 1000 bootstrap replicates were performed to 

obtain branch support (Kamvar et al., 2015). Clone-correction was performed by 

collapsing samples to the average genetic distance detected between technical 

replicates of sequencing. For this, the “mlg.filter” function of the R package Poppr 

(version 2.4.1) was used with a bitwise distance equal to 0.02 and the farthest neighbor 

algorithm. Ordination plots based on principal component analysis (PCA) were also 

constructed to visualize the differentiation of samples within and among subpopulations 

from different host or regions across the Midwest (Fig. 4-1 C). Three distinct PCAs were 

constructed to visualize the relationship among i) all P. cubensis and P. humuli samples 

by host ii) all P. cubensis samples collected from C. sativus and iii) all P. humuli 

samples collected in Michigan and Oregon.  The PCAs were performed using the 

“glPca” function of R package Adegenet (Jombart and Ahmed, 2011) and the ordination 

plots were constructed using the R package Ggplot2. Samples were colored according 

to host or subpopulations and ellipses of 90% confidence intervals were drawn on the 

ordination plots.  

Genetic differentiation between and within subpopulations was quantified using 

hierarchical F-statistics (hierfstat version 0.04.-22, Goudet, 2005) and analysis of 

 

145 

 

molecular variance (AMOVA) (poppr version 2.4.1, Kamvar et al., 2014) in R version 

3.6.1. To test the hypothesis that the Pseudoperonospora spp. populations are 

differentiated by host, we performed an AMOVA and calculated FST statistics using all 

the samples collected by P. cubensis and P. humuli grouped by host. Similarly, to test 

the hypothesis that the P. cubensis subpopulation from C. sativus in Michigan is 

differentiated by geographic location, we performed an AMOVA and calculated FST 

statistics using the samples from the Michigan’s subpopulations (east, west and central) 

and the Midwest-w/o-Mi subpopulation. Population differentiation among the P. humuli 

subpopulations of Michigan (north and west) was analyzed in the same way. A Mantel 

test (Mantel, 1967) was performed to evaluate the relationship between the genetic 

distance of the samples and the physical distance of the location where the samples 

originated. This analysis was performed for each species independently (P. cubensis 

and P. humuli) using only the samples collected in Michigan. The Mantel test was 

performed as described by Gent et al. (2019) using the “mantel.rtest“ function of the 

ade4 package (Dray and Dufour, 2007). The genetic and Euclidean-geographic 

distance matrices were created with the function “bitwise.dist” (Kamvar et al., 2014) and 

“dist”, respectively. Physical distances were calculated using the coordinates of each 

sample location.  

Genotypic diversity. Genotypic diversity was determined using the filtered SNP 

data (not clone corrected) using the R package Poppr (Kamvar et al., 2014). The 

diversity estimates calculated included genotypic richness (MLG), expected number of 

genotypes based on rarefaction (eMLG), evenness (E5) and the indices of Shannon-

Wiener (Shannon, 1948), Stoddart and Taylor’s (Stoddart and Taylor, 1988), and 

 

146 

 

Simpson’s  (Simpson, 1949). The within-population gene diversity (Hexp) (Nei, 1978) 

that describes the proportion of heterozygous genotypes was calculated using the 

function “basic.stats” of the R package hierfstat. The diversity estimates were calculated 

using i) all P. cubensis and P. humuli samples grouped by host ii) all P. cubensis 

samples collected from C. sativus in each Michigan subpopulation and the Midwest-w/o-

Michigan subpopulation iii) the P. humuli samples from the north and west 

subpopulations of Michigan. Only categories or subpopulations with at least five 

individuals were included for this analysis.  

Reproductive Mode. To infer the predominant reproductive mode of P. cubensis 

and P. humuli in Michigan (e.g., sexual, clonal, or mixed) we calculated the index of 

association IA (Brown et al., 1980; Milgroom, 1996) according to Tabima et al. (2018) 

using only the P. cubensis and P. humuli samples collected in the state from C. sativus 

and H. lupulus, respectively. For this, we calculated the index of association (IA) using a 

subset of 1000 random high-quality SNP and compared to the observed value of 

simulated populations with 0, 25, 50, 75, and 100% linkage. The mean IA values of each 

species were compared independently against the mean IA values of the simulated 

populations. Simulations were conducted with a dataset consisting of 2,978 loci 

analogous to the observed P. cubensis and P. humuli data. A nonparametric Kruskal-

Wallis rank-sum test was used to assess the differences among means for all IA 

distributions. Multiple comparisons between IA distribution were performed using the 

nonparametric Kruskal- Wallis test and Tukey’s honest significant difference (HSD) test.  

 

 

147 

 

RESULTS  
 

Library sequencing and read mapping. The sequencing and alignment results 

of all libraries used in this study are summarized in Table 4-1. A total of 270 samples 

containing 2 – 500 ng of DNA were used for library preparation and enrichment. 

However, we could not generate enough enrichment products from 136 libraries, 

including 51 libraries prepared from samples with less than 50 ng of DNA. The 

remaining libraries (N=134) were enriched successfully and were prepared from 

samples containing 2 – 500 ng of DNA; these libraries included 92 libraries of P. 

cubensis and 42 libraries of P. humuli that were sequenced using two lanes of 

HiSeq4000.  

TE libraries: 510 million paired end reads (2 X 150bp) were obtained from the 

sequencing of these libraries (134) generating approximately 3.9 ± 2.6 SD (standard 

deviation) million reads per library. A range of 35 to 54% of the reads aligned to the 

reference genome of P. cubensis. The total number of reads obtained per library was 

influenced by the hybridization time of the libraries with the capture probes for 

enrichment. The libraries that hybridized for 48 h generated almost two times more 

aligned reads (2.64 ± 2.6 SD million) than the libraries that hybridized for 24 h (1.34 ± 

1.0 SD million) (Table 4-1). Less than 20,000 sequencing reads were generated for only 

eight libraries that were eliminated (six samples were hybridized for 24 h and two were 

hybridized for 48 h), leaving 126 enriched libraries for downstream analysis. Five 

percent of the reads per library that did not align to the reference genome of P. cubensis 

aligned to the C. sativus reference genome.  

 

148 

 

Lc-WGS libraries: Seven libraries of P. cubensis and two libraries of P. humuli 

were sequenced without enrichment using MiSeq (Lc-WGS) generating a total of 41 

million sequencing reads (2 X 150 bp). Approximately 0.55 ± 0.37 SD million reads were 

obtained per library, but only 25% of them aligned to the reference genome of P. 

cubensis. Twenty-five percent of the remaining reads aligned to the C. sativus reference 

genome.  

WGS libraries: A total of 350 million paired end reads (2 x 150 bp) were 

obtained from the bioproject PRJNA360426, that included WGS data of 9 libraries of P. 

cubensis and one library of P. humuli. An average of 15.6 ± 6 SD million reads per 

sample we obtained after sequencing. Eighty-nine percent of the total reads aligned to 

the P. cubensis reference genome and approximately 2% of the reads per sample 

aligned to the reference genome of C. sativus.  

Genotyping and variant calling. SNPs were retrieved from the sequencing 

reads generated from 126 enriched DNA libraries and the reads obtained from the 10 

libraries sequenced using whole genome sequencing (Bioproject PRJNA360426). 

Ninety percent of the enriched libraries were prepared from input DNA amounts 

between 2 to 100 ng (Fig. 4-2 B). After mapping the reads to the reference genome of 

P. cubensis, 5,957 biallelic SNPs were identified across 70% of the 136 samples. To 

retain the maximum number of SNPs, we excluded 31 of the samples prepared using 

TE as they contained more than 40% missing data (24 of these were hybridized for 24 

h, with the remaining libraries hybridized for 48 h) (Table 4-1). SNPs with more than 

75% of the missing data across the entire data set were also excluded. This resulted in 

a total of 2,978 high-quality (HQ) SNPs with <10% missing data for downstream 

 

149 

 

analysis. All HQ SNPs were located within 127 out of the 812 genes targeted, 66% of 

the SNPs were found within secreted and effectors proteins. The remaining 44% of the 

SNPs were contained within esterases (16.3%), glucanasas (3%), lyases (3%), SSRs 

(4%), and other (7%) genes.   

The coverage depth (CD) across the 2,978 HQ SNPs was variable among 

samples and significant differences in the CD were observed between the libraries 

genotyped using TE and the libraries genotyped using WGS (Kruskal-Wallis p = 2.2e-

16; Fig. 4-2 A). The enriched libraries with a hybridization time of 48 h (TES3) exhibited 

a significantly higher CD than the enriched libraries generated with 24 h of hybridization 

time (TES2) or the libraries genotyped using WGS (Kruskal-Wallis p = 2.2e-16; Fig. 4-2 

A). The CD across all HQ SNPs had an average of 262 ± 546 SD and 358 ± 546 SD for 

the libraries with hybridization times of 24 and 48 h, respectively (Table 4-1; Fig. 4-2 A). 

Among the samples sequenced using WGS, the CD across HQ SNPs had an average 

of 180 ± 116 SD (Table 4-1). No correlation was observed between the amount of input 

DNA used for library preparation and the CD across HQ SNPs (p > 0.05, Fig. 4-2 B).  

Characterization of technical error. Technical replicates were used to 

characterize the number of errors introduced into a sample during TE. Of the five 

samples with two technical replicates submitted for enrichment, only three samples (18-

35, L32 and L682) and their corresponding replicates were retained after the quality 

filtering of the SNPs. Differences in the average coverage between technical replicates 

of the same sample were observed (Fig. 4-3 A), however, only a small proportion of loci 

were different between replicates (Fig. 4-3 B). On average, 30 SNPs (out of 2,978 total 

 

150 

 

SNPs) differed between technical replicates, suggesting an error rate of roughly 1% 

within the high-quality variants.  

Genetic differentiation and population structure . In the genetic distance tree 

of HQ SNPs (2,978) all 101 samples genotyped fell into one of two clades (100% 

bootstrap support); either P. cubensis (67 samples) or P. humuli (34 samples) (Fig. 4-4). 

Within the P. cubensis clade, samples were either clade I or II of P. cubensis (Thomas 

et al., 2017). The majority of samples from Cucumis spp. aligned with clade II of P. 

cubensis, while the remaining samples collected from C. moschata and C. lanatus 

aligned with clade I (Fig. 4-4).   

Samples in the ordination plot formed 3 clusters corresponding to P. humuli, P. 

cubensis clade I, and P. cubensis clade II; these clusters were consistent with the 

stratification by species and within P. cubensis by host in the genetic distance tree (Fig. 

4-5 A). F-statistics (FST > 0.25) and AMOVA detected significant genetic differentiation 

(P = 0.01) between pathogen species, supporting the clustering of the ordination plot 

(Table 4-3). Stratification by host was supported by a high level of genetic differentiation 

among samples collected from Cucumis spp. and samples collected from C. moschata 

(FST = 0.11, Table 4-3). AMOVA revealed that 78% of the total genetic variance was 

significantly associated with differences between host within P. cubensis (P = 0.01, 

Table 4-4).  

All P. cubensis samples collected from C. sativus in the Midwest including 

Michigan and the reference sample from Florida clustered loosely together across all 

four quadrants of the ordination plot (Fig. 4-5 B). However, the reference samples from 

North Carolina and California clustered separately from each other and were not 

 

151 

 

contained within the 90% confidence ellipses of the other subpopulations (Fig. 4-5 B). 

More than 70% of the samples from the subpopulations of the central and east Michigan 

were contained within the 90% confidence ellipse of the Midwest-w/o-MI subpopulation 

(Fig. 4-5 B); however, only 40% of the samples from west Michigan were contained 

within the ellipse of the of the Midwest-w/o-MI subpopulation. AMOVA did not support 

significant genetic differentiation among the subpopulations of P. cubensis from C. 

sativus in Michigan and the Midwest-w/o-MI subpopulation (P = 0.21). Only 1% of the 

total genetic variance was associated with difference among subpopulations (or 

regions) (Tables 4-6 and 4-7).  

The clone correction of all P. cubensis samples collected from C. sativus (N=57) 

resulted in 40 MLGs, 23% of which were detected in multiple subpopulations (Fig. 4-6 

A, B). MLG 100 was detected in west Michigan and the Midwest-w/o-MI subpopulations 

(Fig. 4-6 A). The subpopulation of west Michigan also shared three MLG with the 

subpopulation of east Michigan (MLG 27, 36 and 46) (Fig. 4-6 A). The subpopulations of 

east and central Michigan shared MLG 33 and MLG 91 and two identical samples 

corresponding to MLG 47 were detected in two distant areas of the Midwest-w/o-MI 

subpopulation (i.e. Indiana and Ontario) (Fig. 4-6 A). Additionally, the MLG 29 from 

Florida was also detected in the east Michigan subpopulation (Fig. 4-6 A). Supporting 

these findings, the Mantel test did not detect a significant relationship between the 

genetic and physical distances of the P. cubensis samples collected from C. sativus in 

the Midwest (P = 0.454, Fig. 4-7 A). 

The level of genetic differentiation among P. humuli samples from Michigan and 

Oregon (Fst = 0.0168) was higher when compared to the genetic differentiation 

 

152 

 

detected among P. cubensis samples from C. sativus originating in the Midwest, 

California, Florida and North Carolina (Fst = 0.0024). In the ordination plot of all the P. 

humuli samples, most samples from the subpopulation of west Michigan were loosely 

dispersed in the left quadrants while most samples from the subpopulation of north 

Michigan (75%) were clustered tightly in the right quadrants (Fig. 4-5 C). Only two 

samples from the north subpopulation were contained within the 90% confidence ellipse 

of the west subpopulation (Fig. 4-5 C). The samples from central and east Michigan 

were either contained in the ellipses from the north or west subpopulations. The 

reference sample from Oregon was not contained within the 90% confidence ellipse of 

Michigan  subpopulations (Fig. 4-5 C). These patterns were supported by AMOVA in 

which significant genetic differentiation between the subpopulations of north and west 

Michigan was detected (P=0.01, Table 4-9). A total of 11.25% of the genetic variance 

was significantly associated with differences among subpopulations and only 2.5% of 

the variance was associated with differences among hop yards within a subpopulation 

(Table 4-10). This suggests that the geographic region has a significant effect on the 

structure of the P. humuli population in Michigan.  

Clone correction of the 34 P. humuli samples resulted in 23 MLGs, only one of 

them (MLG 4) was detected in two different subpopulations (Fig. 4-6 C, D). Only two 

different MLGs were shared among hop yards within regions. MLG 3 was detected in 

two different years at the three yards in north Michigan and MLG 59 was detected once 

in the two hop yards sampled in west Michigan (Fig. 4-6 C). The Mantel test supported 

the hypothesis that the genetic distance among P. humuli samples collected in Michigan 

 

153 

 

was significantly associated (P= 0.0042) with the geographic distance among hop yards 

(Fig. 4-7 B). 

Genotypic diversity. The highest level of genotypic diversity was observed 

among the samples of P. humuli followed by P. cubensis clade I and P. cubensis clade 

II (Table 4-5).  Each species or clade consisted of different MLGs spread relatively 

evenly. This was reflected in the estimates of genotypic richness (eMLG= 6 to 9.54) and 

evenness (E5 > 0.8) (Table 4-5). The Shannon-Wiener index, Stoddart and Taylor’s 

index, and Simpson’s index were highest among P. cubensis clade II samples, followed 

by P. humuli samples, and then P. cubensis clade I samples (Table 4-5). The greatest 

expected heterozygosity or proportion of heterozygous genotypes expected was 

observed among P. humuli samples followed by P. cubensis clade I samples, and then 

P. cubensis clade II samples (Table 4-5). 

 

The Michigan subpopulations and the Midwest-w/o-MI subpopulation of P. 

cubensis collected from C. sativus exhibited similarly low levels of expected 

heterozygosity and genotypic diversity (Table 4-8). Expected heterozygosity ranged 

between 0.178 and 0.197 while genotypic richness (eMLG) ranged between 6 to 9.57. 

The east Michigan subpopulation had the greatest expected heterozygosity (Hexp= 

0.197) followed by Midwest-w/o-MI (Hexp= 0.190), west Michigan (Hexp= 0.189), and 

central Michigan subpopulations (Hexp= 0.178) (Table 4-8). The east Michigan 

subpopulation had the highest genotypic richness (eMLG), while west Michigan and the 

Midwest-w/o-MI subpopulations had larger sample size and lower number of MLGs 

(Table 4-8). The Shannon-Wiener index, Stoddart and Taylor’s index, and Simpson’s 

index were all highest for the west Michigan subpopulation, followed by the 

 

154 

 

subpopulations of east Michigan, Midwest-w/o-MI, and then central Michigan (lowest 

indices) (Table 4-8). The highest evenness (E5) was detected for the central Michigan 

subpopulation followed by the subpopulations of east Michigan, west Michigan and 

Midwest-w/o-MI (lowest E5) (Table 4-8).  

 

The genotypic diversity varied more widely across the P. humuli subpopulations 

of Michigan (eMLG = 5 to 11) compared to the variation detected among P. cubensis 

subpopulations (eMLG = 6 to 9.57). The expected heterozygosity among the 

subpopulations of P. humuli was higher and ranged between 0.206 to 0.220. The 

subpopulations of north (Hexp= 0.224) and east-central Michigan (Hexp= 0.220) had 

the greatest expected heterozygosity followed by the subpopulation of west Michigan 

(Hexp= 0.206) (Table 4-11). The subpopulation of west Michigan had the highest 

genotypic richness, while the north Michigan subpopulation had the largest sample size 

and the most MLGs. The Shannon-Wiener index, Stoddart and Taylor’s index, and 

Simpson’s index were all highest for the north Michigan subpopulation, followed by the 

subpopulations of west and east-central Michigan (lowest indices) (Table 4-11). The 

highest evenness (E5) was detected in the subpopulations of west and east-central 

Michigan followed by the north Michigan subpopulation (lowest E5) (Table 4-11).  

Reproductive Mode. To determine whether allelic variants were randomly 

associated as expected in populations with a sexual mode of reproduction (low linkage), 

we calculated the mean index of association across loci for each species in Michigan 

and compared to the estimated index of association of simulated populations under 

strong linkage (100% linked loci), moderate linkage (75 and 50% linked loci), and low 

linkage (25% and 0 loci under linkage) (Fig. 4-8). Upon comparison, significant 

 

155 

 

differences were observed between the mean index of association (IA) of each species 

and the mean values estimated for the simulated populations (P. cubensis, P < 2.2e-16 

and P. humuli, P < 2.2e-16). The IA mean value of the P. cubensis samples was 

situated between the IA mean values of the simulated data with 50 and 75% linkage 

(Fig. 4-8 A). This indicated that populations of P. cubensis have a mixed-mode of 

reproduction with a predominantly clonal phase. The IA mean value of the P. humuli 

samples was not significantly different to the simulated data with 25% linkage (Fig. 4-8 

B), suggesting that the populations of P. humuli may have a mixed mode of 

reproduction that could be predominantly sexual. 

DISCUSSION 
 

Technological advances in genome sequencing have accelerated the cataloging 

of genomic variation of plant pathogens, however, extracting large amounts of high-

quality DNA from obligate pathogens responsible for DM is challenging, and hampers 

their genotyping using next-generation sequencing. We adapted a TE method that 

facilitated the sequencing of 736 genes annotated as virulence factors in the genome of 

P. cubensis. We used this method to sequence P. cubensis and P. humuli DNA 

extracted from sporangia collected from plant tissue with signs of the pathogen. This 

approach facilitated the genotyping of samples that contained very low amounts of 

pathogen DNA mixed with other environmental contaminants (i.e. plant DNA, bacteria 

DNA). After aligning the sequenced DNA, we identified 2,978 SNPs and resolved the 

population structure of P. cubensis and P. humuli in Michigan. A significant effect of 

location on the genetic variation of P. humuli was detected and the genetic distance 

among samples was associated with the physical distance of the hop yards. Evidence 

 

156 

 

of location-based differentiation within Michigan was not detected for the P. cubensis 

population. 

By using affinity probes and several amplification cycles, our TE protocol was 

designed to provide high coverage sequencing of specific loci in the P. cubensis and P. 

humuli genome. This facilitated the sequencing of a low amount of target (pathogen) 

DNA from environmental samples containing a significant amount of contaminant DNA 

from plant tissues and other microorganisms. Sequencing DNA directly from 

environmental samples reduces time, labor, and DNA input required for other 

sequencing approaches such as genotyping by sequencing (GBS) or WGS (Summers 

et al., 2015b; Thomas et al., 2017a; Gent et al., 2019). This is particularly advantageous 

for DM, as other sequencing approaches (GBS, WGS, Rad-Seq) rely on large amounts 

of high-quality DNA that can only be obtained using a laborious propagation procedure 

(Thomas et al., 2017a; Gent et al., 2019). Using TE, several samples with less than 50 

ng of DNA were successfully genotyped. Most of the samples that failed were leaf 

lesions containing few or no sporangia (P. cubensis had been previously confirmed via 

qPCR). The TE protocol also enabled the genotyping of samples that were no longer 

viable for propagation, expanding the number of samples that could be genotyped 

including those collected more than ten years ago. 

 Using a reduced amount of DNA (2 to 100x less), the samples sequenced after 

TE reached similar sequencing coverage when compared to the samples sequenced 

using a WGS approach, at the region’s target (250x). However, the enriched samples 

required 6 to 7 times less space for sequencing (60 to 70 samples per lane of HiSeq) 

compared to the space used for the sequencing of libraries using WGS (10 samples per 

 

157 

 

lane of HiSeq). This was possible because TE reduced the complexity of the P. 

cubensis genome from 88.22 MB to < 1MB (i.e. 1.13% of the genome). This reduction in 

the genome complexity and the enrichment of up to four samples in a single enrichment 

experiment (MyBaits) make TE a cost-effective alternative to WGS.  

However, we did find that our TE protocol can result in a higher number of 

sequencing errors per sample compared to WGS. The high number of amplification 

cycles used for the enrichment of samples with low amounts of DNA may have led to 

the introduction of PCR errors and a subsequent reduction in genotyping accuracy. 

Typically, high-throughput sequencing (e.g. WGS and GBS) without enrichment of 

samples with low amounts of DNA (<250ng) requires an average of 14 PCR cycles, 

however, 6-14 extra cycles were used after library construction in our TE protocol (for a 

total of 20-28 cycles). Thus, possibly due to the high number of extra amplification 

cycles, the samples genotyped using TE were subject to an error rate of approximately 

1%, an error rate 10 times higher than the estimated error rate for high-throughput 

sequencing data of >0.1% (Grünwald et al., 2017; Ma et al., 2019). 

Despite an error rate of 1%, the genetic distance between samples in 95% of 

cases was greater than the distance generated due to PCR errors, providing confidence 

that our findings reflect patterns linked to the pathogen biology than an artifact of 

sequencing. Using TE, we detected a significant effect of host in the population 

structure of P. cubensis, and the samples collected from C. sativus were genetically 

different from the samples collected from C. moschata. These results support previous 

studies that have shown that the structure of the P. cubensis population is driven by 

host preference, with samples from Cucumis spp. and C. moschata belonging to two 

 

158 

 

distinct evolutionary clades (Thomas et al., 2017a; Wallace et al., 2020). Thomas et al. 

(2017) identified two P. cubensis clades in the U.S.; clade I occurs on C. pepo, C. 

moschata, C.  maxima, C. lanatus, and clade II occurs on Cucumis spp. In our 

analyses, all P. cubensis samples clustered by clades according to that distribution.  

Annual CDM infections in the northern U.S. are driven by an influx of airborne P. 

cubensis sporangia from overwintering sites (Bello et al., 2020; Naegele et al., 2016; 

Goldenhar and Hausbeck, 2019). CDM limits cucumber yield in Michigan where 15000 

ha of cucumbers are planted every year (USDA, 2020). Small, yet significant genetic 

differences among cucumber production regions in Michigan were previously reported 

(Naegele et al., 2016). However, a significant effect of location in the subpopulation 

structure of P. cubensis was not detected in the current study, suggesting no 

differentiation within and among Michigan’s regions. An exchange of migrants has a 

homogenizing effect on subpopulations (Milgroom, 2015) and may occur in Michigan 

due to the availability of susceptible crops across the state. The detection of the same 

multi-locus genotypes (MLG) in multiple locations supports this hypothesis and also 

suggests that P. cubensis sporangia may be disseminated unrestricted in the state as 

there is no geographical barrier (e.g. mountain range or water body) that may limit the 

homogenizing of geographically distant populations. The wide host availability in 

Michigan may also facilitate the establishment of incoming MLG from other 

subpopulations in the Midwest. The exchange of migrants may occur among 

Midwestern states outside of Michigan but the absence of susceptible crops between 

geographically distance populations may result in lower rates of exchange and 

 

159 

 

subsequently more genetically differentiated populations. Additional sampling is needed 

to test this hypothesis.  

Significant differences were detected between the P. humuli populations from the 

north and west Michigan regions, but we did not find any evidence of genetic 

differentiation among hop yards within the same region. In Michigan, approximately 400 

ha of hops are planted across more than 50 commercial hop yards (Michigan 

Department of Agriculture & Rural Development, 2018). A lack of differentiation among 

hop yards within regions could also be a consequence of the continuous exchange of P. 

humuli genotypes among them. However, despite the potential for airborne dispersal of 

hop DM (Bello et al, 2020; Gent et al., 2009) only one MLG was detected in multiple hop 

yards within the same region. This is consistent with the restricted pattern of dispersal 

suggested for this pathogen; new infections of hop plants by P. humuli are less likely to 

occur far from their inoculum source (Johnson et al., 1991). We detected a significant 

correlation between the genetic and geographic distance of P. humuli samples, 

providing circumstantial evidence of a higher probability for new P. humuli infections to 

occur close to their inoculum source.  

In Oregon, Gent et al. (2019) found significant genetic differences between hop 

yards planted within 10 km of each other and a low amount of population differentiation 

between hop yards established from the same planting material. This suggests that 

infected hop plant material may be a more important source of primary inoculum than 

airborne migration for DM. Most of the hop yards sampled for this study were among the 

first yards establish from rhizomes in the mid-2000s when commercial hop production 

returned to Michigan and very few sources of propagation material were available 

 

160 

 

(Sirrine et al., 2014).  As the hop industry matured in the state, new sources of 

propagation material became available, but still in limited numbers. The lack of genetic 

differentiation among hop yards within regions in Michigan may be a result of the 

introduction of a reduced number of genotypes via planting material from the same 

origin source into multiple hop yards. The detection of the same MLGs in multiple hop 

yards and relatively low genetic diversity of north Michigan partially support this 

hypothesis but further sampling and more information on sources of plant material are 

needed to verify this hypothesis in Michigan.  

The restricted airborne dispersal of P. humuli sporangia compare to P. cubensis 

could be attributed to the cultivation practices and geographic distribution of hops. 

Generally, diseased basal shoots are close to the ground and during most of the season 

sheltered within a canopy of healthy basal shoots that hamper dispersal (Johnson et al., 

1991). Additionally, the area planted with hops in the state is equivalent to only 2% of 

the area planted with cucumbers. This lower host availability may result in reduced P. 

humuli sporangia production compared to P. cubensis and lower aerial exchange of 

MLG among subpopulations. The restricted exchange of MLG can result in genetically 

differentiated subpopulations (Milgroom, 2015).  

The relatively low level of genetic differentiation and genetic diversity detected 

within the Pseudoperonospora spp. populations of Michigan is consistent with the clonal 

reproductive mode of P. cubensis and inbreeding reproductive mode of P. humuli 

(Naegele et al., 2016; Gent et al., 2019). Although we estimated relatively low indices 

(65%) of association for both species that suggest they experience a sexual phase, 

especially P. humuli, we believe these values are an artifact of our sampling strategy.  

 

161 

 

Our samples were collected from a bulked inoculum of multiple leaf lesions (cucurbits) 

or diseased shoots (hops), so it is likely that our sample units may contain multiple 

genotypes creating an effect of random mating (no linkage-disequilibrium). An earlier 

study of P. humuli populations in the Pacific Northwest that used a similar sampling 

strategy compared to our study and suggested that the P. humuli population of Oregon 

was reproducing sexually (Chee et al., 2006). However, in a genetic study of P. humuli 

using more precise sample units in the same region, GBS revealed strong evidence of 

linkage disequilibrium (Gent et al., 2019). This is consistent with the nonrandom mating 

expected from species with an asexually reproducing or highly inbreed mode (Gent et 

al., 2019).  Similarly, Wallace et al. (2020) provided evidence of non-random mating or 

recombination consistent with selfing or asexual reproduction for P. cubensis clade II. 

 

Future studies using TE should consider the use of a single lesion as a sample 

unit and sequencing selectively neutral regions (Grünwald et al., 2016). We included 94 

regions (genes) containing small sequence repeats (SSRs) markers among the regions 

targeted for sequencing, as SSRs can provide less biased estimates of population 

differentiation due to their neutrality. However, the sequencing of these regions was of 

poor quality and we could not detect SSRs within them. Instead, we found 115 SNPs 

within these regions that were not significantly differentiated among or within 

subpopulations (data not shown). Most of the signal of population differentiation was 

found within 66% of the HQ SNPs used, these SNPs were contained within secreted 

proteins or effector genes that corresponded to 10% of the genes targeted. Thus, we 

encourage the use of neutral polymorphic regions more evenly distributed across the 

genome.  

 

162 

 

We also recommend the use of a higher number of baits for enrichment when 

working with low concentrations of target DNA and the quantification of plant and 

pathogen DNA using qPCR before library preparation. More baits should increase the 

sequencing coverage across samples, facilitating the calling of variants. Similarly, the 

utilization of samples with a higher amount of pathogen DNA relative to the amount of 

contaminating DNA could increase the chances for the successful preparation of 

libraries using TE. Future studies should also include more technical replicates to 

assess error due to batch effects as the high number of amplification cycles resulted in 

lower genotyping accuracy. The introduction of technical replicates from multiple 

generations of clones is also advised. This is critical when working with high-throughput 

sequencing data due to the introduction of false mutations in the data that could create 

additional multi-locus genotypes (Gent et al., 2019; Potapov and Ong, 2017). 

In summary, our results reveal the key strengths of TE for the genotyping of DM. 

This approach provides a solution for the genotyping of obligate biotrophic pathogens; 

for which high-throughput sequencing is typically constrained by the low amounts of 

target DNA and high amounts of non-target (contaminating) DNA. TE provides a cost-

effective approach for the genotyping of unpurified field samples and the assessment of 

sequence polymorphisms across a large number of individuals. This method could be 

adapted to a diverse group of pathogens even without a reference genome. Future 

population studies using TE should carefully consider the sampling strategy and the 

size and complexity of the genomic regions targeted. Including technical replicates will 

also be important to ensure the accurate genotyping of the samples after enrichment 

and the reproducibility of the experiments. 

 

163 

 

 
 
 
 
 

 
 

 
 
 
 
 
 
 
 
 
 

 

 

 

 

 

 

APPENDIX 

164 

 

 
Table 4-1. Sequencing and alignment results from libraries sequenced using target enrichment (TE), low 
coverage whole genome sequencing (Lc-WGS) and whole-genome sequencing (WGS) 

APPENDIX 

Sequencing characteristics 
Sequencing platform 
Sequencing format (bp) 
Sequencing lines  
Expected sequencing output per lane (Gbp) 
Input amount of DNA for library preparation  
Hybridization time with probes (baits) 
Total number of libraries sequenced 
Total number of clean reads (millions) 
% Reads aligned to reference genome of P. 
cubensis 
% Reads aligned to reference genome of C. sativus 
Aligned reads per samples ± SD (millions) c 
% Libraries with high quality SNPs d 
Average coverage among high quality SNPs ± SD e 

Lc-WGS 

WGSb 

TE seta 1         TE set 2 
HiSeq4000   HiSeq4000   MiSeq.v2.300   HiSeq 
2x100  
2x150  
1 
1 
105 
105 
5-400 ng 
> 1 µg 
-- 
24 h 
10 
62 
211 
350 
89% 
35% 
< 5% 
2% 
16.6 ± 18 
1.34 ± 1.0 
100% 
61% 
262 ± 546 
180 ± 116  

2x150  
2 
3.6 – 4.5 
< 50 ng 
-- 
9 
41 
25% 
> 25% 
0.55 ± 0.37 
-- 
--  

2x150  
1 
105 
5-400 ng 
48 h 
72 
290 
54% 
< 5% 
2.64 ± 2.6 
84% 
358 ± 546  

aTarget enrichment set with hybridization times of 24h and 48h 
bAll the sample genotyped using whole-genome sequencing were retrieved from the bioproject PRJNA360426. 
cMillions of reads aligned to the reference genome of Pseudoperonospora cubensis per sample. 
dLibraries (%) retained containing 2,978 high quality SNPs. 
eAverage coverage among high quality SNPs within the libraries retained. 
 
 
 

 

165 

 

Table 4-2. Plant hosts and the location of the 101 samples of Pseudoperonospora spp. samples used for the 
population analysesa  

West 
Michigan  

East 
Michigan  
 

North 
Michigan  
 

Central 
Michigan  
 

Other states 
within the U.S   Total 
 

 

Pathogen and host species 

P. cubensis clade I 
Cucurbita spp. and Citrullus 
lanatus  
P. cubensis clade II 
Cucumis spp. 
P. humuli 
Humulus lupulus  
 
Totala 

1 

20 
 
8 
 
34 

1 
 
15 
 
0 
 
25 

0 
 
0 
 
20 
 
20 

0 
 
6 
 
5 
 
11 

6 
 
18 
 
1 
 
11 

8 
 
59 
 
34 
 
101 

aThe total of samples reflects the sum of the samples retained after quality filtering.  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

166 

 

Table 4-3. Pairwise FST comparisons among Pseudoperonospora cubensis (clade I and II) and 
Pseudoperonospora humuli. 

FST 
0.11 
0.27 
0.25 

Pvaluea 
0.01 
0.01 
0.01 

Pairwise comparison 
P. cubensis clade II (Cucumis spp.) vs P. cubensis clade I (Cucurbita spp.)  
P. cubensis clade II (Cucumis spp.) vs P. humuli (Humulus lupulus) 
P. cubensis clade I (Cucurbita spp.) vs P. humuli (Humulus lupulus) 

aPvalues were calculated using an analysis of molecular variance (AMOVA) 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

167 

 

Table 4-4. Analysis of molecular variance (AMOVA) for Pseudoperonospora spp. grouped by host. The 
significance of variance was tested from 999 permutations of the dataa. 

MSD d 

Variance (%)e 

Source 
Between host 
Between samples 
within host 
Within samples 
Total  

Sigma 

df b  SSD c 
4033.0104  806.60208  68.370171  78.720821 
5 
2.353212 
301.9755 
2.709468 
30.19755 
10 
16.128063  18.569711 
85 
1370.8853  16.12806 
100  5705.8712  57.05871 
86.851446  100 

Pvalue 
0.01 
 
 
 

aSample sizes used in this analysis included 59, 34 and 8 samples collected from Cucumis spp., Humulus lupulus and 
Cucurbita spp., respectively.   
bDegrees of freedom. 
cSum of squared differences.  
dMean of squared differences. 
eVariance (%) was adjusted to zero for negative sigma values.  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

168 

 

Table 4-5. Genotypic diversity estimates for Pseudoperonospora spp. samples grouped by host (clade). 

Clade/Host species 
P. cubensis clade I 
Cucurbita spp. and Citrullus lanatus 
P. cubensis clade II 
Cucucmis spp. 
P. humuli 
Humulus lupulus  
 
Total 

 

 
1.67 
 

N a  MLG b  eMLG c  SE d  H e 
 
8 
 
59 
 
34 
 
101  77 

 
0 
 
0.634  3.72 
 
0.851  3.17 
 
0.533  4.24 

 
6 
 
9.54 
 
9.14 
 
9.7 

 
6 
 
45 
 
27 
 

 

λ g 
 
0.781 
 

G f  
 
4.57 
 
36.64  0.973 
 
19.27  0.948 
 
59.65  0.983 

 

 

 

Hexp i 
 

E5 h 
 
0.831  0.215 
 
0.889  0.192 
 
0.803  0.217 
 
0.86 

 
0.206 

 

 

aN is the number of individuals sampled of each region.  
bMLG is the number of multilocus genotypes observed.  
ceMLG is the number of expected MLGs at a sample size based on rarefaction.  
dStandard error 
eShannon-Wiener Index (H) 
fStoddart and Taylor’s Index (G) 
gSimpson’s index (lambda) 
hEvenness (E5) 
iWithin population gene diversity (Hexp).  
 
 
 
 
 
 
 
 
 
 
 
 

 

169 

 

Table 4-6. Analysis of molecular variance (AMOVA) of Midwest subpopulationsa of Pseudoperonospora cubensis 
collected from Cucumis sativus. The significance of variance was tested from 999 permutations of the dataa. 

Source 
Between subpopulations 
Between samples within 
subpopulations 
Within samples 
Total 

dfb  SSDc 
3 
35.30 
12  125.89 
38  351.62 
53  512.82 

MSDd  Sigma 
11.76  0.012 
10.49  0.449 
9.25 
9.253 
9.714 
9.67 

% variancee  Pvalue 
0.125 
0.21 
 
4.625 
 
95.248 
 
100 

aSample sizes used in this analysis included 15, 19, and 6 samples collected from Cucumis sativus in the east, west, and 
central Michigan, respectively. The Midwest-w/o-MI subpopulation (17 samples) was formed by all the samples collected 
in the Midwest states of the U.S. and Ontario, Ca, not including Michigan 
bDegrees of freedom. 
cSum of squared differences.  
dMean of squared differences. 
eVariance (%) was adjusted to zero for negative sigma values.  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

170 

 

Table 4-7. Pairwise FST comparisons among subpopulations of Pseudoperonospora cubensis collected from C. 
sativus in the Midwesta. 

Subpopulation 
East Michigan 
Central Michigan  
West Michigan  
Midwest-w/o-Michigan b 
 

East 
Michigan  
-- 
0.0083* 
0.0034 
0.0025 
 

Central 
Michigan  
 
-- 
0.0035 
0.0039 

West 
Michigan  
 
 
-- 
1E-04 

Midwest-w/o-
Michigan 
 
 
 
-- 

aSample sizes used in this analysis included 15, 19, and 6 samples collected from Cucumis sativus in the east, west, and 
central Michigan, respectively. 
bThe Midwest-w/o-MI subpopulation (17 samples) was formed by all the samples collected in the Midwest states of the 
U.S. and Ontario, Ca, not including Michigan 
*Indicates a significant Fst value based on AMOVA 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

171 

 

Table 4-8. Genotypic diversity estimates and index of association of Pseudoperonospora cubensis 
subpopulations  collected from C. sativus in the Midwest. 

Subpopulation 
East Michigan 
Central Michigan 
West Michigan 
Midwest-w/o-Michiganj   
Total 

N a  
15 
6 
19 
17 
57 

MLG b  eMLG c  SE d 
14 
0.495 
0 
6 
0.772 
16 
0.795 
14 
44 
0.638 

9.57 
6 
9.07 
8.85 
9.54 

H e 
2.62 
1.79 
2.7 
2.56 
3.69 

G f 
13.2 
6 
13.4 
11.6 
35.7 

λ g 
0.924 
0.833 
0.925 
0.913 
0.972 

E5 h 
0.965 
1 
0.893 
0.887 
0.886 

Hexp i 
0.197 
0.178 
0.189 
0.19 
0.192 

aN is the number of individuals sampled of each region.  
bMLG is the number of multilocus genotypes observed.  
ceMLG is the number of expected MLGs at a sample size based on rarefaction.  
dStandard error 
eShannon-Wiener Index (H) 
fStoddart and Taylor’s Index (G) 
gSimpson’s index (lambda) 
hEvenness (E5) 
iWithin population gene diversity (Hexp).  
jThe Midwest-w/o-MI subpopulation was formed by all the samples collected in the Midwest states of the U.S. and 
Ontario, Ca, not including Michigan 
   
 
 
 
 
 
 
 
 
 
 
 

 

172 

 

Table 4-9. Analysis of molecular variance (AMOVA) of Michigan subpopulations of Pseudoperonospora humuli. 
The significance of variance was tested from 999 permutations of the dataa 

Source 
Between subpopulations 
Between samples within 
subpopulations 
Within samples 
Total  

Df b  SSD c  MSD d 
1 

40.187  40.18681  1.986518  11.249556 

Sigma  % variance e  Pvalue 
0.01 
 

3 
23 
27 

53.042  17.68075  0.449694  2.546595 
350.12  15.22243  15.22243  86.203849 
443.34  16.42018  17.65864  100 

 
  

aSample sizes used in this analysis included 8 and 20 samples from two and three hop yards in west and north Michigan, 
respectively.   
bDegrees of freedom. 
cSum of squared differences.  
dMean of squared differences. 
eVariance (%) was adjusted to zero for negative sigma values.  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

173 

 

Table 4-10. Pairwise FST comparisons among subpopulations of Pseudoperonospora humuli collected from hop 
yards in Michigana. 

Subpopulation/Yard 

West  

North  

D 
E 
B 
A 
C 

West  
D 
-- 
0.00237 
0.02269* 
0.03052* 
0.02879* 

North 
B 
 
 
-- 

E 
 
-- 
0.00953 
0.021636*  0.002614 
0.019606*  0.001831 

A 
 
 
 
-- 
0.002613 

C 
 
 
 
 
-- 

aSample sizes used in this analysis included 4, 4, 6, 4 and 10 samples collected from hop yards indicated as D, E, C, B 
and A, respectively. The hop yards in the central region were excluded from analysis due to a low number of samples. 
*Indicates a significant Fst value based on AMOVA 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

174 

 

Table 4-11.Genotypic diversity estimates and index of association of Michigan subpopulations of 
Pseudoperonospora humuli. 

E5 h 

λ g 
0.875  1 
2.08  8 
0.83 
2.11  5.88 
1.61  5 
0.8 
2.86  12.52  0.92 

0.676 
1 
0.701 

Hexp i 
0.206 
0.224 
0.220 
0.217 

8 
11 
5 
22 

8 
6.79 
5 
8.33 

Subpopulation 
West  
North  
East-central 
Total 

N a  MLG b  eMLG c  SE d  H e   G f 
8 
20 
5 
33 

0 
1.05 
0 
1.08 
aN is the number of individuals sampled of each region.  
bMLG is the number of multilocus genotypes observed.  
ceMLG is the number of expected MLGs at a sample size based on rarefaction.  
dStandard error 
eShannon-Wiener Index (H) 
fStoddart and Taylor’s Index (G) 
gSimpson’s index (lambda) 
hEvenness (E5) 
iWithin population gene diversity (Hexp).  
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

175 

 

 
Figure 4-1. Cucumis sativus planted acreage and number Humulus lupulus planted in Michigan by county 
(Adapted from Neufeld, 2017).  

 

A)  C. sativus planted acreage, location and number of samples collected in the Michigan (N=40). A total of 26 
samples were collected outside of Michigan in Ontario Canada (N=6) and the U.S. states of Florida (N=2), 
Alabama (N=1), North Carolina (N=2), South Carolina (N=2), California (N=1), New York (N=1), Wisconsin (N=1), 
Ohio (N=3), Iowa (N=1), Indiana (N=5), Georgia (N=1). (B) H. lupulus planted, location and number of samples 
collected from hop yards in Michigan (N=33) and Oregon (N=1). Hop yards with more than 800 plants (H. lupulus) 
are represented by stars, those with fewer are represented by circles. Hop yard sampled (A-F) are colored in red. 
(C) Population assignment by geographic location of Pseudoperonospora spp. samples in the Midwest (Wisconsin, 
Indiana, Ohio, Michigan (Mi) and Ontario, CA). 

176 

 
 
 

 

 

A

t

)
h
p
e
d

 

e
g
a
r
e
v
o
C

(
 

0
1
g
o

l

3

2

1

R = -0.053 , p = 0.78
R = 0.077 , p = 0.59

Tec

TES2
TES3
WGS

B

3.5

)
h

t

 

p
e
d
e
g
a
r
e
v
o
C

(
 
0
1
g
o

l

3.0

2.5

2.0

1.5

1.0

set
a
a
a

s2
s3
wgs

2
S
E
T

3
S
E
T

Sequencing approach

S
G
W

0

25

50

ng of DNA

75

100

Figure 4-2. Depth coverage of high-quality SNPs across libraries.  
(A). Depth coverage of high-quality SNPs (log10) across libraries of Pseudoperonospora cubensis and 
Pseudoperonospora humuli sequenced using target enrichment with 24h (TES2) and 48h (TES3) of hybridization and 
whole genome sequencing (WGS). (B). Depth coverage of high-quality SNPs (log10) across libraries sequenced using 
target enrichment with different amounts of input DNA for library preparation. Only samples that were retained after quality 
filtering were included in the analysis.  
 
 
 
 
 
 
 

 

 

177 

A
5
3
8
1

.

B
5
3
8
1

.

 

A
2
3

B
2
3

Sample

A
3
7
6

0.00

B
3
7
6

A
2
8
6

B
2
8
6

0.02

0.00
0.04

0.02
0.06

18−35A S37

0.04
0.08

0.06

Genetic distance (proportion of loci that are different) 2683 SNPs

Genetic distance (proportion of loci that are different) 2683 SNPs

A
2
8
6

B
2
8
6

A

B

Genetic distance

3

2

1

)
h
)
t
h
p
t
p
e
e
d
d
 
 
e
e
g
g
a
a
r
e
r
v
e
o
v
C
o
(
 
c
0
(
1
0
g
1
o
g
o
L

l

Tec

TES2
TES3

18-35B_S39
L32_S62
L682_S63

100

18-35A_S37 L32_S26 L682_S28
0.169
0.022
0.0104

0.176
0.011
0.174

0.011
0.176
0.173

100

100

100

100

L32 S62

L32 S26

L682 S63

L682 S28

L673A S4

100

18−35B S39

18−35A S37

7
3
S
_
A
5
3
−
8
1

9
3
S
_
B
5
3
−
8
1

6
2
S
_
2
3
L

2
6
S
_
2
3
L

Sample
Sample

4
S
_
A
3
7
6
L

8
2
S
_
2
8
6
L

3
6
S
_
2
8
6
L

0.00
Genetic distance (proportion of loci that are different (2978 SNPs))

Genetic distance. Proportion of loci that are different (2978 SNPs)

0.04

0.02

0.06

0.08

Figure 4-3. Characterization of sequencing error among technical replicates.  
(A) Coverage distribution of the technical replicates sequenced. Violin plots are filled with 2,978 high-quality SNPs. Each 
sample is colored according to the sequencing batch. (B) Genetic differentiation among technical replicates. A UPGMA 
tree was reconstructed using 2,978 SNPs. The genetic distance represents the number of SNPs that are different among 
samples.

 

 

178 

 

 

Cucurbita.spp
Cucurbita.spp
Cucurbita.spp

C.sativus
C.sativus
C.sativus

H.lupulus
H.lupulus
H.lupulus

Cucurbita.spp
Cucurbita.spp
Cucurbita.spp
Cucurbita moschata
Cucumis sativus
Humulus lupulus
C.sativus
C.sativus
C.sativus
Citrullus lanatus
Cucurbita maxima
Cucumis melo
H.lupulus
H.lupulus
H.lupulus

86

100

100

100

93

100

100

86

100

84

78

Pseudoperonospora

cubensis clade I

98

77

100

99

413 S36

***
**

39
24
37
41
38
10D S8
L682 S63
L682 S28
L1335 S16
44
898 S15
156 S51
1484 S9
1755 S4
CDM207 S55
257 S50
CDM19 S22
L33 S3
1752 S32
L119 S27
L1621 S13
L1072 S21
CDM209 S11
CDM201 S9
CDM232 S59
238 S11
CDM154 S60
936 S35
1589 S7
L32 S62
L32 S26
L1755 S64
1649 S2
L573 S20
204 S10
CDM228 S6
CDM202 S12
253 S49
1335 S30
K2A S38
CDM191 S14
186 S3
CDM250 S17
200 S5
180 S14
CDM242 S23
14J S6
OH1−2 S20
229 S8
203 S1
L1024 S15
481 S13
CDM155 S31
L673A S4
CDM177 S25
413 S36
413 S36
CDM152 S30
CDM123 S7
671 S34
CDM110 S48
S17−1 S17
CPF7 S24
296 S12
CDM153 S29
42
40
40
44
44
43
18−25 S36
17−15 S33
18−23 S37
17−101 S74
18−32 S64
17 108B S41
18−26 S55
18−17 S52
17−8 S50
17−20 S76
17−112 S18
17−100 S73
18−31 S65
17−113 S72
A19 S21
18−7 S66
18−20 S54
18−6 S67
kk10−4 S75
17−103 S70
18−44 S59
18−42 S42
18−5 S56
18−15 S60
18−1 S69
18−43 S44
17−114 S71
58 S61
18−30 S53
18−41 S45
18−35B S39
18−35A S37
18−13 S27
25

***
*****

100

93

100

100

100

39
24
37
41
38
10D S8
L682 S63
L682 S28
L1335 S16
898 S15
156 S51
1484 S9
1755 S4
CDM207 S55
257 S50
CDM19 S22
L33 S3
84
1752 S32
L119 S27
L1621 S13
L1072 S21
CDM209 S11
78
CDM201 S9
CDM232 S59
238 S11
CDM154 S60
936 S35
1589 S7
L32 S62
L32 S26
L1755 S64
1649 S2
L573 S20
204 S10
CDM228 S6
90
39
CDM202 S12
24
253 S49
37
1335 S30
41
K2A S38
38
CDM191 S14
10D S8
186 S3
L682 S63
CDM250 S17
L682 S28
200 S5
L1335 S16
180 S14
898 S15
CDM242 S23
98
156 S51
14J S6
1484 S9
OH1−2 S20
1755 S4
229 S8
CDM207 S55
203 S1
257 S50
L1024 S15
73
CDM19 S22
481 S13
L33 S3
CDM155 S31
1752 S32
L673A S4
L119 S27
CDM177 S25
L1621 S13
413 S36
L1072 S21
CDM152 S30
CDM209 S11
CDM123 S7
CDM201 S9
671 S34
100
CDM232 S59
CDM110 S48
238 S11
S17−1 S17
CDM154 S60
CPF7 S24
936 S35
296 S12
1589 S7
CDM153 S29
L32 S62
42
100
L32 S26
40
L1755 S64
44
100
1649 S2
43
L573 S20
18−25 S36
204 S10
17−15 S33
CDM228 S6
18−23 S37
CDM202 S12
17−101 S74
253 S49
18−32 S64
1335 S30
17 108B S41
K2A S38
18−26 S55
CDM191 S14
18−17 S52
186 S3
17−8 S50
82
CDM250 S17
17−20 S76
200 S5
17−112 S18
180 S14
17−100 S73
CDM242 S23
18−31 S65
97
14J S6
17−113 S72
OH1−2 S20
A19 S21
229 S8
18−7 S66
71
203 S1
18−20 S54
L1024 S15
18−6 S67
481 S13
kk10−4 S75
CDM155 S31
17−103 S70
L673A S4
18−44 S59
93
CDM177 S25
18−42 S42
413 S36
18−5 S56
83
CDM152 S30
18−15 S60
83
CDM123 S7
18−1 S69
73
671 S34
18−43 S44
CDM110 S48
17−114 S71
S17−1 S17
58 S61
CPF7 S24
74
18−30 S53
296 S12
18−41 S45
CDM153 S29
18−35B S39
42
18−35A S37
40
18−13 S27
44
25
43
18−25 S36
17−15 S33
18−23 S37
17−101 S74
18−32 S64
17 108B S41
18−26 S55
18−17 S52
17−8 S50
17−20 S76
17−112 S18
17−100 S73
18−31 S65
17−113 S72
A19 S21
18−7 S66
18−20 S54
18−6 S67
kk10−4 S75
17−103 S70
18−44 S59
18−42 S42
18−5 S56
18−15 S60
18−1 S69
18−43 S44
17−114 S71
58 S61
18−30 S53
18−41 S45
18−35B S39
18−35A S37
18−13 S27
25

0.00

100

Cucurbita.spp
Cucurbita.spp
Cucurbita.spp

C.sativus
C.sativus
C.sativus

H.lupulus
H.lupulus
H.lupulus

C.lanatus
C.lanatus
C.lanatus

C. maxima
C. maxima
C. maxima

C. melo
C. melo
C. melo

100

100

100

100

100

100

73
84

86

100

93

100

100

100

100
100

90

100

100
90

100
84

78

78

100

86

100

100

100

90

93

100

100

100

100

100

C. melo
C. melo
C. melo

C. melo
C. melo
C. melo

C.sativus
C.sativus
C.sativus

H.lupulus
H.lupulus
H.lupulus

C.lanatus
C.lanatus
C.lanatus

C.lanatus
C.lanatus
C.lanatus

C.lanatus
C.lanatus
C.lanatus

C. maxima
C. maxima
C. maxima

C. maxima
C. maxima
C. maxima

C. maxima
C. maxima
C. maxima

Cucurbita.spp
Cucurbita.spp
Cucurbita.spp

cubensis clade II

Pseudoperonospora 

39
24
37
41
38
10D S8
L682 S63
L682 S28
L1335 S16
898 S15
156 S51
1484 S9
1755 S4
CDM207 S55
257 S50
CDM19 S22
L33 S3
1752 S32
L119 S27
L1621 S13
98
L1072 S21
CDM209 S11
CDM201 S9
CDM232 S59
238 S11
CDM154 S60
936 S35
1589 S7
L32 S62
L32 S26
L1755 S64
1649 S2
L573 S20
204 S10
CDM228 S6
CDM202 S12
253 S49
1335 S30
K2A S38
CDM191 S14
186 S3
CDM250 S17
200 S5
180 S14
CDM242 S23
14J S6
OH1−2 S20
229 S8
203 S1
L1024 S15
481 S13
82
CDM155 S31
L673A S4
80
CDM177 S25
413 S36
97
CDM152 S30
CDM123 S7
671 S34
CDM110 S48
S17−1 S17
CPF7 S24
296 S12
CDM153 S29
42
40
44
43
18−25 S36
17−15 S33
18−23 S37
17−101 S74
18−32 S64
17 108B S41
18−26 S55
18−17 S52
100
17−8 S50
17−20 S76
17−112 S18
17−100 S73
Figure 4-4. Genetic differentiation of Pseudoperonospora cubensis and 
18−31 S65
17−113 S72
A19 S21
Pseudoperonospora humuli samples.  
18−7 S66
18−20 S54
0.02
18−6 S67
kk10−4 S75
17−103 S70
A UPGMA tree was reconstructed using 2,978 single nucleotide polymorphism (SNP) 
18−44 S59
80
18−42 S42
variants. Bootstrap support values are indicated above the branches. The genetic 
18−5 S56
18−15 S60
18−1 S69
distance represents the proportion of loci that are different between samples. * Samples 
18−43 S44
71
17−114 S71
58 S61
previously classified as clade I members. ** Samples previously classified as clade II 
18−30 S53
18−41 S45
members. Technical replicates are enclosed in red squares. 
18−35B S39
18−35A S37
18−13 S27
25

Genetic distance  (proportion of loci that are different)

Pseudoperonospora 

Genetic distance (proportion of loci that are different)
Genetic distance (proportion of loci that are different)
Genetic distance (proportion of loci that are different)

Genetic distance (proportion of loci that are different)
Genetic distance (proportion of loci that are different)
Genetic distance (proportion of loci that are different)

humuli

0.02

100
100

100

82

97

82

97

100

100

100

83

83

74

100

100

100

C. melo
C. melo
C. melo

0.00

98

93

83

71

73

71

73

98

73

0.04

80

73

100

74

100

80

100

93

83

100

 

100

100

93

83

73

83

 

179 

74

100

Genetic distance (proportion of loci that are different)
Genetic distance (proportion of loci that are different)
Genetic distance (proportion of loci that are different)

66

60

74

Genetic distance (proportion of loci that are different) 2683 SNPs

Genetic distance (proportion of loci that are different) 2683 SNPs

0.06

 

Subpopulation

Subpopulation

Figure 4-5. Ordination plots of Pseudoperonospora spp. based on 2,978 SNPs. 
((A) Ordination plot of Pseudoperonospora cubensis and P. humuli samples according 
to host species. All points represent samples collected in Michigan unless indicated 
otherwise. (B) Ordination plot of P. cubensis samples from Cucumis sativus from 2007 
to 2017 in the U.S. (n=57). (C) Ordination plot of P. humuli samples from hop yards of 
Michigan in 2017 and 2018 (n=34) 

 

 

180 

 

 
A

C

D

North

North

West

Central

B

East

N = 57
MLG = 40

West

Central

Midwest-w/o-Mi

Figure 4-6. Frequency and geographic distribution of Pseudoperonospora cubensis and Pseudoperonospora 
humuli genotypes (MLG). 

 

(A) Geographic distribution of P. cubensis genotypes collected from Cucumis sativus in the Midwest. Unique 

genotypes are colored in gray. (B) Frequency of P. cubensis genotypes collected from C. sativus in the Midwest. 
Unique genotypes are colored in gray. (C) Geographic distribution of P. humuli genotypes collected from Humulus 
lupulus in Michigan. Unique genotypes are color coded in gray. The circles and stars colored in red represent 

 

181 

 

different hop yards. (D) Frequency of P. humuli genotypes collected from H. lupulus in Michigan. Unique 
genotypes are colored in gray.  

 
Figure 4-7. The relationship between genetic differences among samples and geographic distances of the 
locations from which samples originated.  

(A) P. cubensis samples from Cucumis sativus and (B) P. humuli samples from Humulus lupulus in Michigan. 

 

 

182 

e

B

0.12

c

e

d

 

A

0.125

0.100

a

0.075

0.050

d

a

c

)

A

i

I
(
 
n
o
i
t
a
c
o
s
s
a
 
f
o
x
e
d
n

 

I

0.025

0.000

b

e
g
a
k
n

i
l

_
o
N

s

i

s
n
e
b
u
c

 
.

P

e
g
a
k
n

e
g
a
k
n

e
g
a
k
n

i
l

_
%
5
2

i
l

_
%
0
5

i
l

_
%
5
7

i

n
o
i
t
a
c
o
s
s
a
 
f
o
x
e
d
n

 

I

e
g
a
k
n

i
l

_
%
0
0
1

0.08

a

0.04

0.00

i
l

u
m
u
h

 
.

P

b

e
g
a
k
n

i
l

_
o
N

a

e
g
a
k
n

e
g
a
k
n

e
g
a
k
n

i
l

_
%
5
2

i
l

_
%
0
5

i
l

_
%
5
7

e
g
a
k
n

i
l

_
%
0
0
1

 

Data set

Figure 4-8. Estimation of the degree of linkage disequilibrium within Pseudoperonospora cubensis and 
Pseudoperonospora humuli of Michigan.  
(A) Observed index of association (IA) distribution of P. cubensis samples compared with the IA distribution values of 0, 
25, 50, 75, and 100% linkage. (B) Observed IA distribution of P. humuli samples compared with the IA distribution values 
of 0, 25, 50, 75, and 100% linkage. Groupings based on the Kruskal-Wallis rank-sum test are noted by the letters over the 
boxplots, in which the P. cubensis and the P. humuli population datasets are grouped with the simulated 50% and 25% 
linkage data, respectively

Data set

 

183 

 

SUPPLEMENTARY TABLES 
 
Supplementary Table 4-1:  Samples sequenced using target enrichment (TE), whole genome sequencing (WGS) or low 
coverage whole genome sequencing (Lc-WGS) 
 

Hybr. 
Timea  

Enri. 
setb 

Log10 
Depth 

State 

County / 

Yard 

Region   Year  

Host 
Species   Collector 

Sample 

10D_S8 

Input 
DNA 
(ng) 
2 

48 h  ES3 

1.9  Michigan  Newaygo 

Eastern 

2015 

1335_S30 

83 

24 h  ES2 

1.6  Michigan 

St. Clair 

Eastern 

2009  C. sativus 

C. 

moschata   J. Bello  
L, 
Quesada-
Ocampo, 
L, 
Quesada-
Ocampo, 

2009  C. sativus 

2015  C. sativus  J, Bello 
2017  C. sativus  J, Bello 

2009  C. sativus 

L, 
Quesada-
Ocampo, 
L, 
Quesada-
Ocampo, 
2017  H. lupulus  D, Higgins 
2017  H. lupulus  D, Higgins 

2009  C. sativus 

Leelanau 

(B) 

(A) 

Northern 

2017  H. lupulus  D, Higgins 

1484_S9 

14J_S6 
156_S51 

1589_S7 

8 

4 
14 

17 

48 h  ES3 

1.75  Michigan  Newaygo  Western 

48 h  ES3 
48 h  ES3 

2.08  Michigan  Newaygo  Western 
1.49  Michigan  Muskegon  Western 

48 h  ES3 

2.69 

Indiana 

1649_S2 

54 

48 h  ES3 

2.46 

Iowa 

17-100_S73 
17-101_S74 

6 
5 

48 h  ES3 
48 h  ES3 

2.05  Michigan  Barry (D)  Western 
1.34  Michigan 
Northern 

Leelanau 

- 

- 

Midwest 

Midwest 

17-103_S70 

54 

48 h  ES3 

2.52  Michigan 

 
aHybridization time. bEnrichment set (ES2 = 24h, ES3=48h) 
 
 

 

184 

Sample 

17_108B_S41 

17-112_S18 

17-113_S72 

17-114_S71 

17-15_S33 

17-20_S76 

1752_S32 

1755_S4 

17-8_S50 

180_S14 

Input 
DNA 
(ng) 
54 

85 

54 

34 

38 

5 

72 

23 

25 

4 

Hybr. 
Timea  

Enri. 
setb 

48 h 

ES3 

48 h 

ES3 

State 

County / 

Log10 
Depth 
Yard 
1.28  Michigan  Genesee 
Leelanau 
2.08  Michigan 

(F) 

48 h 

ES3 

2.33  Michigan 

48 h 

ES3 

2.49  Michigan 

48 h 

ES3 

1.45  Michigan 

Leelanau 

Leelanau 

Leelanau 

(A) 

(A) 

(A) 

(B) 

24 h 

ES2 

1.86  Michigan  Berrien (E)  Western 

2017 

48 h 

ES3 

1.96  Michigan 

Ingham  

Central 

2009 

48 h 

ES3 

24 h 

ES2 

48 h 

ES3 

1.61  Michigan 
Ingham  
1.45  Michigan  Genesee 
2.84  Michigan  Muskegon  Western 

Central 

Central 

(F) 

2009 

2017 

2018 

Region   Year  

Host 
Species   Collector 

Central 

2017 

Northern 

2017 

Northern 

2017 

Northern 

2017 

Northern 

2017 

H. 

H. 

H. 

H. 

H. 

lupulus  D, Higgins 
lupulus  D, Higgins 
lupulus  D, Higgins 
lupulus  D, Higgins 
lupulus  D, Higgins 
lupulus  D, Higgins 
L, Quesada-
Ocampo, 
L, Quesada-
Ocampo, 
sativus 
lupulus  D, Higgins 

sativus 

H. 

C. 

H. 

C. 

C. 

sativus 

J, Bello 

 

Supplementary Table 4-1. (cont’d) 
 

 
aHybridization time. bEnrichment set (ES2 = 24h, ES3=48h) 
 
 
 
 
 
 

 

185 

 

Sample 

18-13_S27 

Input 
DNA 
(ng) 
32 

18-15_S60 

136 

18-17_S52 

18-1_S69 

18-20_S54 

18-23_S37 

33 

45 

26 

56 

18-25_S36 

145 

18-26_S55 

18-30_S53 

30 

42 

18-31_S65 

205 

48 h 

48 h 

48 h 

48 h 

48 h 

24 h 

24 h 

48 h 

48 h 

48 h 

ES3 

ES3 

ES3 

ES3 

ES3 

ES2 

ES2 

ES3 

ES3 

ES3 

Grand 

2.84  Michigan 

2.72  Michigan 

Traverse (C)  Northern 
Traverse (C)  Northern 
Traverse (C)  Northern 
Traverse (C)  Northern 
1.56  Michigan  Genesee (F)  Central 

1.63  Michigan 

Michigan 

Grand 

Grand 

3 

2018 

2017 

2017 

2018 

2017 

1.08  Michigan  Berrien (E)  Western 

2018 

1.43  Michigan  Berrien (E)  Western 

2018 

1.51  Michigan 

Barry (D)  Western 

2018 

2.95  Michigan  Leelanau (A)  Northern 

2018 

2.28  Michigan  Leelanau (A)  Northern 

2018 

H. 

H. 

H. 

H. 

H. 

H. 

H. 

H. 

H. 

H. 

lupulus 

lupulus 

lupulus 

lupulus 

lupulus 

lupulus 

lupulus 

lupulus 

lupulus 

lupulus 

J, Bello 
D, 
Higgins 
D, 
Higgins 
J, Bello 
D, 
Higgins 
J, Bello 

J, Bello 

J, Bello 

J, Bello 

J, Bello 

Supplementary Table 4-1. (cont’d) 
 

Hybr. 
Timea  

Enri. 
setb 

Log10 
Depth 

State 

Region   Year  

Host 
Species   Collector 

County / 

Yard 
Grand 

 
aHybridization time. bEnrichment set (ES2 = 24h, ES3=48h) 
 
 
 
 
 
 

 

186 

 

Supplementary Table 4-1. (cont’d) 
 

Sample 

18-32_S64 

18-35A_S37 

18-35B_S39 

18-41_S45 

18-42_S42 

18-43_S44 

18-44_S59 

18-5_S56 

186_S3 

18-6_S67 

Input 
DNA 
(ng) 
342 

54 

54 

37 

62 

26 

20 

41 

26 

37 

Hybr. 
Timea  

Enri. 
setb 

Log10 
Depth 

State 

County / 

Yard 

Region   Year  

Host 
Species   Collector 

48 h 

48 h 

48 h 

48 h 

48 h 

48 h 

48 h 

48 h 

48 h 

48 h 

ES3 

ES3 

ES3 

ES3 

ES3 

ES3 

ES3 

ES3 

ES3 

ES3 

1.3 

Michigan  Barry (D)  Western  2018 

3.02  Michigan 

3.04  Michigan 

2.79  Michigan 

2.97  Michigan 

2.94  Michigan 

2.99  Michigan 

2.98  Michigan 

Leelanau 

Leelanau 

(A) 

(A) 

(A) 

(B) 

(B) 

(B) 

(A) 

Leelanau 

Leelanau 

Leelanau 

Leelanau 

Leelanau 

Northern  2018 

Northern  2018 

Northern  2018 

Northern  2018 

Northern  2018 

Northern  2018 

Northern  2018 

2.39  Michigan  Muskegon  Western  2018 

2.07  Michigan  Berrien (E)  Western  2018 

lupulus 

lupulus 

lupulus 

lupulus 

lupulus 

lupulus 

lupulus 

lupulus 

H. 

H. 

H. 

H. 

H. 

H. 

H. 

H. 

C. 

H. 

sativus 

lupulus 

J, Bello 

J, Bello 

J, Bello 

J, Bello 

J, Bello 

J, Bello 

J, Bello 

J, Bello 

J, Bello 

J, Bello 

 
aHybridization time. bEnrichment set (ES2 = 24h, ES3=48h) 
 
 
 
 
 
 

 

187 

Supplementary Table 4-1. (cont’d) 
 

Input 
DNA 
(ng) 
392 

Hybr. 
Timea  

Enri. 
setb 

Log10 
Depth 

State 

48 h 

ES3 

1.76  Michigan 

County 
/ Yard 
Barry 
(D) 

Region   Year  

Host 
Species   Collector 

Western  2018 

 

Sample 

18-7_S66 

200_S5 

203_S1 

204_S10 

229_S8 

238_S11 

lupulus 

sativus 

sativus 

sativus 

sativus 

sativus 

H. 

C. 

C. 

C. 

C. 

C. 

C. 

a  
C. 

C. 

moschat

sativus 

J, Bello 

J, Bello 

J, Bello 

J, Bello 

J, Bello 

J, Bello 

Thomas, A 

J, Bello 

J, Bello 

sativus 
lupulus  Thomas, A 

H. 

20 

74 

8 

14 

7 

48 h 

ES3 

2.94  Michigan  Saginaw  Eastern 

2018 

48 h 

ES3 

1.71  Michigan  Saginaw  Eastern 

2018 

48 h 

ES3 

2.59  Michigan  Saginaw  Eastern 

2018 

48 h 

ES3 

2.29  Michigan  Berrien   Western  2018 

48 h 

ES3 

1.87  Michigan  Allegan  Western  2018 

24  NA 

wgs 

wgs 

2.33 

Alabama 

- 

Alabama  2013 

253_S49 

257_S50 

7 

3 

48 h 

ES3 

2.8 

Michigan  Saginaw  Eastern 

2018 

48 h 

ES3 

2.12  Michigan  Saginaw  Eastern 

2018 

25  NA 

wgs 

wgs 

2.11 

Oregon 

- 

Oregon 

2012 

 
aHybridization time. bEnrichment set (ES2 = 24h, ES3=48h) 
 
 
 
 
 

 

188 

 

Supplementary Table 4-1. (cont’d) 
 

Sample 

296_S12 

Input 
DNA 
(ng) 
5 
37  NA 

Enri. 
Hybr. 
Timea  
setb 
48 h  ES3 
wgs  wgs 

State 

Log10 
Depth 
1.56  Michigan 
2.38 

County 
/ Yard 
Ingham  

- 

-  

- 

- 

Central 
South 
Carolina 
Florida  
South 
Carolina 
North 
Carolina 

South 
Carolina 
Florida  
South 
Carolina 
North 
Carolina 

38  NA 

wgs  wgs 

2.43 

39  NA 

wgs  wgs 

2.3 

40  NA 

wgs  wgs 

1.97 

Region  

Year  

Host 
Species   Collector 

C. 

C. 

moschata  

2018  C. sativus  J, Bello 
Thomas, 
2013 
A 
2013  C. lanatus  Thomas, 
A 
Thomas, 
2012 
A 
2013  C. maxima  Thomas, 

moschata  

moschata  

2007  C. melo 

A 
L, 
Quesada-
Ocampo, 
Thomas, 
A 
L, 
2008  C. sativus 
Quesada-
Ocampo, 
2012  C. sativus  Thomas, 
A 
Thomas, 
2013  C. melo 
A 

2008 

C. 

413_S36 

26 

48 h  ES3 

1.32  Michigan  Clinton 

Eastern 

41  NA 

wgs  wgs 

1.93 

Georgia 

42  NA 

wgs  wgs 

1.97  California 

43  NA 

wgs  wgs 

44  NA 

wgs  wgs 

2.33 

North 
Carolina 
2.42  New York 

- 

- 

- 

- 

Georgia 

California 

Ncarolina 

New York 

 
aHybridization time. bEnrichment set (ES2 = 24h, ES3=48h) 
 
 
 
 
 

 

189 

Hybr. 
Timea  

Enri. 
setb 

Log10 
Depth 

State 

County / 

Yard 

Region   Year  

Host 
Species   Collector 

481_S13 

5 

48 h 

ES3 

0.9 

Indiana 

- 

Midwest  2007  C. sativus 

58_S61 

197 

48 h 

ES3 

3.07  Michigan 

Ingham 

(G) 

Central 

L, 
Quesada-
Ocampo, 

2017  H. lupulus  D, 
Higgins 
L, 
Quesada-
Ocampo, 
L, 
Quesada-
Ocampo, 
L, 
Quesada-
Ocampo, 

Midwest  2008  C. sativus 

Midwest  2008  C. sativus 

Midwest  2008  C. sativus 

48 h 
24 h 
24 h 
24 h 
24 h 

ES3 
ES2 
ES2 
ES2 
ES2 

2019  H. lupulus  J, Bello 
2.41  Michigan 
2017  C. sativus  J, Bello 
1.2  Michigan 
1.11  Michigan 
2017  C. sativus  J, Bello 
1.18  Michigan  Muskegon  Western  2017  C. sativus  J, Bello 
1.52  Michigan  Muskegon  Western  2017  C. sativus  J, Bello 

Ingham   Central 
Ingham   Central 
Ingham   Central 

 

Supplementary Table 4-1. (cont’d) 
 

Sample 

Input 
DNA 
(ng) 

48 h 

ES3 

1.3 

Ontario 

48 h 

ES3 

1.92 

Ohio 

48 h 

ES3 

2.63  Ontario 

- 

- 

- 

671_S34 

898_S15 

936_S35 

A19_S21 
CDM110_S48 
CDM123_S7 
CDM152_S30 
CDM153_S29 

7 

2 

7 

67 
16 
10 
8 
12 

 
aHybridization time. bEnrichment set (ES2 = 24h, ES3=48h) 
 
 
 
 
 
 

 

190 

 

Supplementary Table 4-1. (cont’d) 
 

Sample 

Input 
DNA 
(ng) 

Hybr. 
Timea  

Enri. 
setb 

Log10 
Depth 

State 

County 
/ Yard 

Region   Year  

Host 
Species   Collector 

L32_S26 

10 

48 h 

ES3 

2.43  Michigan  Monroe  Eastern 

2007  C. sativus 

L32_S62 

10 

24 h 

ES2 

2.61  Michigan  Monroe  Eastern 

2008  C. sativus 

L33_S3 

31 

24 h 

ES2 

1.8 

Michigan  Monroe  Eastern 

2007  C. sativus 

L573_S20 

5 

24 h 

ES2 

2.47 

Florida  

-  

Florida  

2008  C. sativus 

L673A_S4 

20 

24 h 

ES2 

1.08 

Ontario 

L682_S28 

22 

24 h 

ES2 

NA 

Ontario 

L682_S63 

11 

24 h 

ES2 

2.1 

Ontario 

OH1-2_S20 
S17-1_S17 

55 
100 

48 h 
48 h 

ES3 
ES3 

1.9 
1.11 

Ohio 
Indiana 

 
aHybridization time. bEnrichment set (ES2 = 24h, ES3=48h) 
 
 

- 

- 

- 

- 
- 

 

191 

L, 
Quesada-
Ocampo, 
L, 
Quesada-
Ocampo, 
L, 
Quesada-
Ocampo, 
L, 
Quesada-
Ocampo, 
L, 
Quesada-
Ocampo, 
L, 
Quesada-
Ocampo, 
L, 
Quesada-
Ocampo, 

Midwest 

2008  C. sativus 

Midwest 

2008  C. sativus 

Midwest 

2008  C. sativus 

Midwest 
Western 

2017  C. sativus  J, Bello 
moschata   J, Bello 
2017 

C. 

 

Supplementary Table 4-1. (cont’d) 
 

Sample 

CDM232_S59 

Input 
DNA 
(ng) 
26 

CDM242_S23 

CDM250_S17 

18 

10 

Hybr. 
Timea  

Enri. 
setb 

Log10 
Depth 

State 

County / 

Yard 

Region   Year  

Host 
Species   Collector 

24 h 

ES2 

2.35  Michigan 

Allegan  Western 

2018 

24 h 

ES2 

3.4 

Michigan 

Berrien   Western 

2018 

24 h 

ES2 

2.95  Michigan 

Berrien   Western 

2018 

sativus 

sativus 

sativus 

C. 

C. 

C. 

C. 

C. 

H. 

sativus 

sativus 

lupulus 

C. 

sativus 

C. 

sativus 

C. 

sativus 

C. 

sativus 

J, Bello 

J, Bello 

J, Bello 

J, Bello 

J, Bello 

J, Bello 
L, 
Quesada-
Ocampo, 
L, 
Quesada-
Ocampo, 
L, 
Quesada-
Ocampo, 
L, 
Quesada-
Ocampo, 

CPF7_S24 

162 

48 h 

ES3 

1.11 

Indiana 

- 

Midwest 

2017 

K2A_S38 

kk10-4_S75 

L1024_S15 

L1072_S21 

32 

15 

6 

8 

48 h 

ES3 

2.41  Michigan 

Bay 

Eastern 

2016 

48 h 

ES3 

2.25  Michigan 

Leelanau 

(B) 

Northern 

2018 

24 h 

ES2 

1 

Indiana 

- 

Midwest 

2008 

24 h 

ES2 

2.02  Michigan  Monroe  

Eastern 

2008 

L119_S27 

22 

24 h 

ES2 

1.74 

Ohio 

- 

Midwest 

2007 

L1335_S16 

2 

48 h 

ES3 

1.86  Michigan 

St. Clair 

Eastern 

2009 

 
aHybridization time. bEnrichment set (ES2 = 24h, ES3=48h) 
 
 

 

192 

 

Supplementary Table 4-1. (cont’d) 
 

Hybr. 
Timea  

Enri. 
setb 

Log10 
Depth 

State 

County / 

Yard 

Region   Year  

Host 
Species   Collector 

sativus 

sativus 

sativus 

C. 

C. 

C. 

C. 

C. 

C. 

C. 

C. 

C. 

sativus 

sativus 

sativus 

sativus 

sativus 

sativus 

C. 

sativus 

C. 

sativus 

J, Bello 

J, Bello 

J, Bello 

J, Bello 

J, Bello 

J, Bello 

J, Bello 

J, Bello 

J, Bello 
L, 
Quesada-
Ocampo, 
L, 
Quesada-
Ocampo, 

Sample 

CDM154_S60 

CDM155_S31 

CDM191_S14 

CDM19_S22 

CDM201_S9 

CDM202_S12 

CDM207_S55 

CDM209_S11 

CDM228_S6 

L1621_S13 

Input 
DNA 
(ng) 

44 

32 

24 

83 

52 

56 

92 

28 

25 

40 

24 h 

24 h 

24 h 

24 h 

24 h 

24 h 

24 h 

24 h 

24 h 

ES2 

ES2 

ES2 

ES2 

ES2 

ES2 

ES2 

ES2 

ES2 

1.74  Michigan  Muskegon  Western 

2017 

1 

Michigan  Muskegon  Western 

2017 

2.75  Michigan  Muskegon  Western 

2018 

1.85 

Ontario 

- 

Midwest 

2016 

2.28  Michigan  Saginaw 

Eastern 

2018 

2.67  Michigan  Saginaw 

Eastern 

2018 

1.93  Michigan 

Berrien   Western 

2018 

2.35  Michigan 

Berrien   Western 

2018 

2.58  Michigan 

Berrien   Western 

2018 

24 h 

ES2 

2.16  Wisconsin 

- 

Midwest 

2009 

L1755_S64 

22 

24 h 

ES2 

2.57  Michigan 

Ingham  

Central 

2009 

 
aHybridization time. bEnrichment set (ES2 = 24h, ES3=48h) 

 

193 

 

 

 

 

 

 

 

 

 

 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

LITERATURE CITED 

194 

 

LITERATURE CITED 

 
 
Ali, S., Gautier, A., Leconte, M., Enjalbert, J., de Vallavieille-Pope, C. (2011) A rapid 
genotyping method for an obligate fungal pathogen, Puccinia striiformis f.sp. tritici, 
based on DNA extraction from infected leaf and multiplex PCR genotyping. BMC Res 
Notes 4: 240 
 
Baxter, L., Tripathy, S., Ishaque, N., Boot, N., Cabral, A., Kemen, E., Thines, M., Ah-
Fong, A., Anderson, R., Badejoko, W., et al (2010) Signatures of adaptation to obligate 
biotrophy in the Hyaloperonospora arabidopsidis genome. Science (80) 330: 1549 LP – 
1551 
 
Bello, J.C., Sakalidis, M.L., Perla, D.E., Hausbeck, M.K. (2020) Detection of airborne 
sporangia of Pseudoperonospora cubensis in Michigan using Burkard spore traps 
coupled to qPCR. Plant Dis. - Manuscript in review 
 
Blum, M., Waldner, M., Olaya, G., Cohen, Y., Gisi, U., Sierotzki, H. (2011) Resistance 
mechanism to carboxylic acid amide fungicides in the cucurbit downy mildew 
pathogen Pseudoperonospora cubensis. Pest Manag Sci 67: 1211–1214 
 
Bolger, A.M., Lohse, M., Usadel, B. (2014) Trimmomatic: a flexible trimmer for Illumina 
sequence data. Bioinformatics 30: 2114–2120 
 
Bourret, T.B., Choudhury, R.A., Mehl, H.K., Blomquist, C.L., McRoberts, N., Rizzo, D.M. 
(2018) Multiple origins of downy mildews and mito-nuclear discordance within the 
paraphyletic genus Phytophthora. PLoS One 13: e0192502 
 
Brown, A.H.D., Feldman, M.W., Nevo, E. (1980) Multilocus structure of natural 
populations of Hordeum spontaneum. Genetics 96: 523-536. 
 
Burkhardt, A., Buchanan, A., Cumbie, J..S, Savory, E.A., Chang, J.H., Day, B. (2014) 
Alternative splicing in the obligate biotrophic oomycete pathogen Pseudoperonospora 
cubensis. Mol Plant-Microbe Interact 28: 298–309 
 
Burkhardt, A., Day, B. (2016) Transcriptome and small RNAome dynamics during a 
resistant and susceptible interaction between cucumber and downy mildew. Plant 
Genome. doi: 10.3835/plantgenome2015.08.0069 
 
Carleson, N.C., Fieland, V.J., Scagel, C.F., Weiland, J.E., Grünwald, N.J. (2019) 
Population structure of Phytophthora plurivora on rhododendron in Oregon nurseries. 
Plant Dis 103: 1923–1930 
 
Chee, H.Y., Nelson, M.E., Grov,e G.G., Eastwell, K.C., Kenny, S.T., Klein, R.E. (2006) 
Population biology of Pseudoperonospora humuli in Oregon and Washington. Plant Dis 
90: 1283–1286 

 

195 

 

 
Choi, Y-J., Hong, S-B., Shin, H-D. (2005) A re-consideration of Pseudoperonospora 
cubensis and P. humuli based on molecular and morphological data. Mycol Res 109: 
841–848 
 

Choi, Y-J., Thines, M. (2015) Host jumps and radiation, not co-divergence drives 

diversification of obligate pathogens. A case study in Downy mildews and Asteraceae. 
PLoS One 10: e0133655 
 
Clark, M.J., Chen, R, Lam, H.Y.K., Karczewski, K.J., Chen, R., Euskirchen, G., Butte, 
A.J., Snyder, M. (2011) Performance comparison of exome DNA sequencing 
technologies. Nat Biotechnol 29: 908–914 
 
Cruz-Dávalos, D.I., Llamas, B., Gaunitz, C., Fages, A., Gamba, C., Soubrier, J., 
Librado, P., Seguin-Orlando, A., Pruvost, M., Alfarhan, A.H., et al (2017) Experimental 
conditions improving in-solution target enrichment for ancient DNA. Mol Ecol Resour 17: 
508–522 
 
Cui, C., Herlihy, J.H., Bombarely, A., McDowell, J.M., Haak, D.C. (2019) Draft assembly 
of Phytophthora capsici from long-read sequencing uncovers complexity. Mol Plant-
Microbe Interact 32: 1559–1563 
 
Dray, S., Dufour, A. (2007) The ade4 package: implementing the duality diagram for 
ecologists. J Stat Softw 22(4): 1–20 
 
Faircloth, B.C., Branstetter, M.G., White, N.D., Brady, S.G. (2015) Target enrichment of 
ultraconserved elements from arthropods provides a genomic perspective on 
relationships among Hymenoptera. Mol Ecol Resour 15: 489–501 
 
Fletcher, K., Gil, J., Bertier, L.D., Kenefick, A., Wood, K.J., Zhang, L., Reyes-Chin-Wo, 
S., Cavanaugh, K., Tsuchida, C., Wong, J., et al (2019) Genomic signatures of 
heterokaryosis in the oomycete pathogen Bremia lactucae. Nat Commun 10: 2645 
 
Gent, D.H., Adair, N., Knaus, B.J., Grünwald, N.J. (2019) Genotyping-by-sequencing 
reveals fine-scale population differentiation in populations of Pseudoperonospora 
humuli. Phytopathology. doi: 10.1094/PHYTO-12-18-0485-R 
 
Gent, D.H., Nelson, M.E., Farnsworth, J.L., Grove, G.G. (2009) PCR detection 
of Pseudoperonospora humuli in air samples from hop yards. Plant Pathol 58: 1081–
1091 
 
Gent, D.H., Ocamb, C.M., Farnsworth, J.L. (2010) Forecasting and management of hop 
downy mildew. Plant Dis 94: 425–431 
 
Gisi, U., Sierotzki, H. (2008) Fungicide modes of action and resistance in downy 
mildews. Eur J Plant Pathol 122: 157–167 

 

196 

 

 
Gnirke, A., Melnikov, A., Maguire, J., Rogov, P., LeProust., E. M., Brockman, W., 
Fennell, T., Giannoukos, G., Fisher, S., Russ, C., et al (2009) Solution hybrid selection 
with ultra-long oligonucleotides for massively parallel targeted sequencing. Nat 
Biotechnol 27: 182–189 
 
Goldenhar, K.E., Hausbeck, M.K. (2019) Fungicides for control of downy mildew on 
pickling cucumber in Michigan. Plant Heal Prog 20: 165–169 
 
Goudet, J. (2005) Hierfstat, a package for r to compute and test hierarchical F-statistics. 
Mol Ecol Notes 5: 184–186 
 
Grünwald, N.J., Everhart, S.E., Knaus, B.J., Kamvar, Z.N. (2017) Best practices for 
population genetic analyses. Phytopathology 107: 1000–1010 
 
Grünwald, N.J., McDonald, B.A., Milgroom, M.G. (2016) Population genomics of fungal 
and oomycete pathogens. Annu Rev Phytopathol 54: 323–346 
 
Holdsworth, W.L., Summers, C.F., Glos, M., Smart, C.D., Mazourek, M. (2014) 
Development of downy mildew-resistant cucumbers for late-season production in the 
Northeastern United States. HortScience horts 49: 10–17 
 
Johnson, D.A, Alldredge, J.R, Allen, J.R, Allwine, R. (1991) Spatial pattern of downy 
mildew in hop yards during severe and mild disease epidemics. Phytopathology 81: 
1369–1374 
 
Jombart, T., and Ahmed, I. (2011) adegenet 1.3-1: new tools for the analysis of 
genome-wide SNP data. Bioinformatics 27: 3070–3071 
 
Jones, L., Riaz, S., Morales-Cruz, A., Amrine, K.C.H., McGuire, B., Gubler, W.D., 
Walker, M.A., Cantu, D. (2014) Adaptive genomic structural variation in the grape 
powdery mildew pathogen, Erysiphe necator. BMC Genomics 15: 1081 
 
Jouet, A., Saunders, D.G.O., McMullan, M., Ward, B., Furzer, O., Jupe, F., Cevik, V., 
Hein, I., Thilliez, G.J.A., Holub, E., van Oosterhout, C. and Jones, J.D.G. (2019), Albugo 

candida race diversity, ploidy, and host-associated microbes revealed using DNA 

sequence capture on diseased plants in the field. New Phytol, 221: 1529-1543 
 
Kamvar, Z.N., Brooks, J.C., Grünwald, N.J. (2015) Novel R tools for analysis of 
genome-wide population genetic data with emphasis on clonality. Front Genet 6: 208 
 
Kamvar, Z.N., Tabima, J.F., Grünwald, N.J. (2014) Poppr: an R package for genetic 
analysis of populations with clonal, partially clonal, and/or sexual reproduction. PeerJ 2: 
e281 
 

 

197 

 

Kanetis, L., Vasiliou, A., Neophytou, G., Samouel, S., Tsaltas, D. (2013) First report of 
downy mildew caused by Peronospora belbahrii on sweet basil (Ocimum basilicum) in 
Cyprus. Plant Dis 98: 283 
 
Kitner, M., Lebeda, A., Sharma, R., Runge, F., Dvořák, P., Tahir, A., Choi, Y-J., 
Sedláková, B., Thines, M. (2015) Coincidence of virulence shifts and population genetic 
changes of Pseudoperonospora cubensis in the Czech Republic. Plant Pathol 64: 
1461–1470 
 
Knaus, B.J., Grünwald, N.J. (2017) vcfr: a package to manipulate and visualize variant 
call format data in R. Mol Ecol Resour 17: 44–53 
 
Kozarewa, I., Armisen, J., Gardner, A.F., Slatko, B.E., Hendrickson, C.L. (2015) 
Overview of target enrichment strategies. Curr Protoc Mol Biol 112: 7.21.1-7.21.23 
 
Kunjeti, S.G., Anchieta, A., Martin, F.N., Choi, Y-J., Thines, M., Michelmore, R.W., 
Koike, S.T., Tsuchida, C., Mahaffee, W., Subbarao, K.V., et al (2016) Detection and 
quantification of Bremia lactucae by spore trapping and quantitative PCR. 
Phytopathology 106: 1426–1437 
 
Lane, C.R., Beales, P.A., O’Neill, T.M., McPherson, G.M., Finlay, A.R., David, J., 
Constantinescu, O., Henricot, B. (2005) First report of Impatiens downy mildew 
(Plasmopara obducens) in the UK. Plant Pathol 54: 243 
 
Li, H., Durbin, R. (2009) Fast and accurate short read alignment with Burrows-Wheeler 
transform. Bioinformatics. doi: 10.1093/bioinformatics/btp324 
 
Li, H., Handsaker, B., Wysoker, A., Fennell, T., Ruan J., Homer, N., Marth, G., 
Abecasis, G., Durbin, R., Subgroup 1000 genome project data processing (2009) The 
sequence alignment/map format and SAMtools. Bioinformatics 25: 2078–2079 
 
Lim, H.C., Braun, M.J. (2016) High-throughput SNP genotyping of historical and modern 
samples of five bird species via sequence capture of ultraconserved elements. Mol Ecol 
Resour 16: 1204–1223 
 
Ma, X., Shao, Y., Tian, L., Flasch, D.A., Mulder, H.L., Edmonson, M.N., Liu, Y., Chen, 
X., Newman, S., Nakitandwe, J., et al (2019) Analysis of error profiles in deep next-
generation sequencing data. Genome Biol 20: 50 
 
Mantel, N. (1967) The detection of disease clustering and a generalized regression 
approach. Cancer Res 27: 209–220 
 
McCormack, J.E., Tsai, W.L.E., Faircloth, B.C. (2016) Sequence capture of 
ultraconserved elements from bird museum specimens. Mol Ecol Resour 16: 1189–
1203 
 

 

198 

 

Milgroom, M.G. (2015) Population biology of plant pathogens : genetics, ecology, and 
evolution. The American Phytopathological Society, St. Paul, Minnesota 
 
Milgroom, M.G. (1996) Recombination and the multilocus structure of fungal 
populations. Annu Rev Phytopathol 34: 457–477 
 
Mitchell, M.N., Ocamb, C.M., Grünwald, N.J., Mancino, L.E., Gent, D.H. (2011) Genetic 
and pathogenic relatedness of Pseudoperonospora cubensis and P. 
humuli. Phytopathology 101: 805–818 
 
Naegele, R.P., Quesada-Ocampo, L.M., Kurjan, J.D., Saude, C., Hausbeck, M.K. 
(2016) Regional and temporal population structure of Pseudoperonospora cubensis in 
Michigan and Ontario. Phytopathology 106: 372–379 
 
Nei, M (1972) Genetic distance between populations. Am Nat 106: 283–292 
 
Neufeld, K. N. (2017) Epidemiology of cucurbit downy mildew: predicting risk of disease 
infection and validation of the CDM ipmPIPE forecasting system. (Doctoral dissertation 
thesis, North Carolina State University, Raleigh, USA). Retrieved from 
https://repository.lib.ncsu.edu/bitstream/handle/1840.20/33592/etd.pdf?sequence=1&is
Allowed=y 
 
Perea, C., De La Hoz, J.F., Cruz, D.F., Lobaton, J.D., Izquierdo, P., Quintero, J.C., 
Raatz, B., Duitama, J. (2016) Bioinformatic analysis of genotype by sequencing (GBS) 
data with NGSEP. BMC Genomics 17: 498 
 
Potapov, V., Ong, J.L. (2017) Examining sources of error in PCR by single-molecule 
sequencing. PLoS One 12: e0169774 
 
Purayannur, S., Cano, L.M., Bowman, M.J., Childs, K.L., Gent, D.H., Quesada-Ocampo, 
L.M. (2020) The effector repertoire of the hop downy mildew pathogen 
Pseudoperonospora humuli. Front Genet   11: 910 
 
Quesada-Ocampo, L.M., Granke, L.L., Olsen, J., Gutting, H.C., Runge, F., Thines, M., 
Lebeda, A., Hausbeck, M.K. (2012) The genetic structure of Pseudoperonospora 
cubensis populations. Plant Dis 96: 1459–1470 
 
Quinlan, AR, Hall, IM (2010) BEDTools: a flexible suite of utilities for comparing 
genomic features. Bioinformatics 26: 841–842 
 
Rahman, A., Góngora-Castillo, E., Bowman, M.J., Childs, K.L., Gent, D.H., Martin, F.N., 
Quesada-Ocampo, L.M. (2019) Genome sequencing and transcriptome analysis of the 
hop downy mildew pathogen Pseudoperonospora humuli reveal species-specific genes 
for molecular detection. Phytopathology 109: 1354–1366 
 

 

199 

 

Rivera, Y., Salgado-Salazar, C., Gulya, T.J., Crouch, J.A. (2016) Newly emerged 
populations of Plasmopara halstedii infecting rudbeckia exhibit unique genotypic profiles 
and are distinct from sunflower-infecting strains. Phytopathology 106: 752–761 
 
Savory, E.A., Adhikari, B.N., Hamilton, J.P., Vaillancourt, B., Buell, C.R., Day, B. 
(2012a) mRNA-Seq Analysis of the Pseudoperonospora cubensis transcriptome during 
cucumber (Cucumis sativus L.) infection. PLoS One 7: e35796 
 
Savory, E.A., Zou, C., Adhikari, B.N., Hamilton, J.P., Buell, C.R., Shiu, S-H., Day, B. 
(2012b) Alternative splicing of a multi-drug transporter from Pseudoperonospora 
cubensis generates an RXLR effector protein that elicits a rapid cell death. PLoS One 7: 
e34701 
 
Simpson, E. H. (1949) Measurement of diversity. Nature 163: 688 
 
Sirrine, R., Lizotte, E., Brown, D., O’Brien, T., and Leach, A. 2014. Estimated costs of 
producing hops in Michigan. Michigan State University Extension E3236. Online 
publication. 
 
Shannon, C. E. (1948). A mathematical theory of communication. Bell System Technical 
Journal 27:379-423 
 
Sharma, R., Xia, X., Cano, L.M., Evangelisti, E., Kemen, E., Judelson, H., Oome, S., 
Sambles, C., van den Hoogen, D.J., Kitner, M., et al (2015) Genome analyses of the 
sunflower pathogen Plasmopara halstedii provide insights into effector evolution in 
downy mildews and Phytophthora. BMC Genomics 16: 741 
 
Starrett, J., Derkarabetian, S., Hedin, M., Bryson, Jr. R.W., McCormack, J.E., Faircloth, 
B.C. (2017) High phylogenetic utility of an ultraconserved element probe set designed 
for Arachnida. Mol Ecol Resour 17: 812–823 
 
Stassen, J., Seidl, M., Vergeer, P., Nijman, I., Snel, B., Cuppen, E., Van den 
ackerveken, G. (2012) Effector identification in the lettuce downy mildew Bremia 
lactucae by massively parallel transcriptome sequencing. Mol Plant Pathol 13: 719–731 
 
Stoddart, J.A., Taylor, J.F. (1988) Genotypic diversity: estimation and prediction in 
samples. Genetics 118: 705 LP – 711 
 
Summers C.F., Gulliford C.M., Carlson C.H., Lillis J.A., Carlson M.O., Cadle-Davidson 
L., Gent, D.H., Smart, C.D. (2015) Identification of genetic variation between obligate 
plant pathogens Pseudoperonospora cubensis and P. humuli using RNA sequencing 
and genotyping-by-sequencing. PLoS One 10: e0143665 
 
Tabima, J.F., Coffey, M.D., Zazada, I.A., Grünwald, N.J. (2018) Populations 
of Phytophthora rubi show little differentiation and high rates of migration among states 
in the western United States. Mol Plant-Microbe Interact 31: 614–622 

 

200 

 

 
Thines, M. (2014) Phylogeny and evolution of plant pathogenic oomycetes—a global 
overview. Eur J Plant Pathol 138: 431–447 
 

Thilliez, G.J.A., Armstrong, M.R., Lim, T.-Y., Baker, K., Jouet, A., Ward, B., van 

Oosterhout, C., Jones, J.D.G., Huitema, E., Birch, P.R.J. and Hein, I. (2019), Pathogen 
enrichment sequencing (PenSeq) enables population genomic studies in oomycetes. 
New Phytol, 221: 1634.-1648 
 
Thomas, A., Carbone, I., Choe, K., Quesada-Ocampo, L.M., Ojiambo, P.S. (2017) 
Resurgence of cucurbit downy mildew in the United States: Insights from comparative 
genomic analysis of Pseudoperonospora cubensis. Ecol Evol 7: 6231–6246 
 
USDA (2020) United States Department of Agriculture. 2019. Vegetables Annual 
Summary.  
 
Wallace, E.C., D’Arcangelo, K.N., Quesada-Ocampo, L.M. (2020) Population analyses 
reveal two host-adapted clades of Pseudoperonospora cubensis, the causal agent of 
cucurbit downy mildew, on commercial and wild cucurbits. Phytopathology. doi: 
10.1094/PHYTO-01-20-0009-R 
 
Wallace, E.C., Quesada-Ocampo, L.M. (2017) Analysis of microsatellites from the 
transcriptome of downy mildew pathogens and their application for characterization 
of Pseudoperonospora populations. PeerJ 5: e3266 
 
Withers, S., Gongora-Castillo, E., Gent, D., Thomas, A., Ojiambo, P.S., Quesada-
Ocampo, L.M. (2016) Using next-generation sequencing to develop molecular 
diagnostics for Pseudoperonospora cubensis, the cucurbit downy mildew pathogen. 
Phytopathology 106: 1105–1116 
 
Wong, F.P, and Wilcox, W.F (2001) Comparative physical modes of action of 
azoxystrobin, mancozeb, and metalaxyl against Plasmopara viticola (Grapevine Downy 
Mildew). Plant Dis 85: 649–656 
 
 
 

 

 

 

 

 

 

201 

 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

 

 

 

 

 

 

 

 

CHAPTER V 

202 

 

CONCLUSIONS 
 

The research reported in this dissertation provides an assessment of the 

population structure of Pseudoperonospora cubensis, the causal agent of cucurbit 

downy mildew (CDM) in Michigan and evaluates the performance of spore traps 

coupled to qPCR for the monitoring of airborne Pseudoperonospora spp. sporangia. 

The population structure of P. cubensis was investigated using a target enrichment 

protocol that allowed the genotyping of environmental samples with low concentrations 

of a mix of plant and pathogen DNA. A significant effect of the host type on the 

population structure of P. cubensis was observed while no evidence of location-based 

differentiation was detected within the P. cubensis population of Michigan.  In addition, 

this study identified an improved detection system for the monitoring of P. cubensis 

sporangia that allowed the differentiation between Pseudoperonospora spp. and the 

detection of P. cubensis DNA before symptoms were observed in commercial cucumber 

fields. 

This study provided evidence of significant genetic differentiation among the P. 

cubensis population from squash (clade I) and cucumber (clade II) but there was 

insufficient evidence to conclude that location (region) within Michigan has a significant 

effect on the distribution of the genetic variation of the P. cubensis population. 

Contrasting evidence was found for the population of Pseudoperonospora humuli (the 

causal agent of hop downy mildew), and a significant effect of location on the genetic 

variation of the population was detected in the state. The differences in the distribution 

of genetic variation between the population of each species could be explained by 

differences in the dispersion patterns between them. Although both pathogens 

 

203 

 

propagate via asexual sporangia that are aerially dispersed, empirical evidence 

indicates that only P. cubensis sporangia may spread unrestricted over long distances. 

The isolation and divergence of geographically distant populations of Michigan may be 

limited by the exchange of migrants between them. This was supported by the detection 

of the same multilocus genotypes of P. cubensis in multiple geographically distant 

populations. On the other hand, despite the potential for airborne dispersal of hop 

downy mildew, only one MLG was detected in two geographically distant populations 

which is consistent with the restricted pattern of dispersion suggested for this pathogen. 

This was also supported by a significant correlation between the genetic and 

geographic distance among P. humuli samples, that was not detected for the P. 

cubensis population. 

In Michigan, CDM occurs annually due to an influx of aerially dispersed P. 

cubensis sporangia from overwinter sources. Thus, timely alerts of an influx of the 

airborne inoculum of P. cubensis can assist Michigan growers in assessing the need to 

initiate fungicide sprays. In the current study, we reported the use of a highly specific 

and sensitive qPCR assay that allows the differentiation between P. humuli, and each 

host-adapted clade of P. cubensis (clade I and II) on spore trap samples. A distinction 

that was not possible using light microscopy only. After two years of monitoring using a 

Burkard and impaction spore traps coupled with qPCR in cucumber fields, P. humuli 

DNA was detected more frequently than P. cubensis early during the growing season 

from May to June. P. cubensis clade II DNA was detected in spore trap samples 

approximately 2-7 days before CDM symptoms were observed in commercial cucumber 

fields in July or August, while P. cubensis clade I DNA was never detected. The 

 

204 

 

differences in the airborne inoculum of each clade documented here are likely the result 

of differences in the total area planted of the susceptible hosts to each clade in 

Michigan.  

In addition, this study sheds light on the utilization of Burkard and impaction 

spore traps for the airborne monitoring or Pseudoperonospora spp. sporangia. In 

agreement with theoretical expectations, our results suggest that the Burkard spore 

traps are a more efficient instrument for the detection of airborne sporangia at low 

concentrations (<100 sporangia/day) than impaction spore traps. Adjustments can be 

made to increase the efficiency of the detection of P. cubensis using impaction spore 

traps. This includes the utilization of multiple impaction spore traps per location and 

increasing the sampling surface width and the exposition time of the impaction rods. 

The use of spore traps couple with qPCR could be used as part of a CDM risk advisory 

system to time fungicide applications that protect cucurbit crops in Michigan. 

 

 

205