IDENTIFICATION OF VIRAL CONTAMINATION ON LETTUCE FROM THE FIELD TO POST - HARVEST PROCESSING By Samantha Lynn Wengert A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Fisheries and Wildlife - Master of Science 201 5 ABSTRACT IDENTIFICATION OF VIRAL CONTAMINATION ON LETTUCE FROM THE FIELD TO POST - HARVEST PROCESSING By Samantha Lynn Wengert Viral f oodborne outbreaks are a serious threat to public health and fresh produce is becoming increasingly recognized as a transmission vehicle . Potential pre - and post - harvest sources of contamination include irrigation and processing water, soil, manure, equipment, and human handling. T raditional detection methods limit studies on vir uses in produce . New culture - independent metagenomic next generation sequencing (NGS) technologies present an opportunity for generating an improved understanding of the virus communities (virome) associated with foods. The goals of this study were to use NGS techn ology for the first time to identify the virome present in irrigation water and lettuce in the field and to i nvestigate the efficacy of current post - harvest leafy green processing and disinfection practices during a contamination event . In this study, most viruses found in irrigation water and lettuce from the field environment could not be identified suggesting limited knowledge of the vir ome in these environments. H uman enteric viruses such as rotavirus A and picobirnavirus were identified in field lettuce. On the processing side, the efficacy of a chlorine - based sanitizer against MS2 coliphage on fresh - cut romaine lettuce was assessed during simulated commercial production of fresh - cut lettuce . Flume w ashing lettuce in 25 ppm of free chlorin e did not significantly reduce viral levels on romaine lett uce when compared to water without chlorine . Overall, this study suggests that metagenomic technology can be used as a potential tool for monitoring food safety. Viruses were present in field lett uce and resistant to current commercial chlorine disinfection techniques, posing a po ssible threat to public health. Copyright by SAMANTHA LYNN WENGERT 2015 iv ACKNOWLEDGEMENTS This work was supported by a grant from the United Stat es Department of Agriculture, National Institute for Food and Agriculture Grant # 1128137 0 Proposal 2012 - 04389 . I would like to thank my main advisor, Dr. Joan B. Rose, for all her support and encouragement throughout my time as a graduate student at MSU. I would also like to thank Tiong G. Aw, Rebecca Ives, and other Rose Lab members for their help and support in carrying out this research project. Thank you to my committee members Dr. Mohamad Faisal and Dr. Elliot Ryser for their guidance, with specia l recognition to Dr. Elliot Ryser who provided the small - scale leafy green processor for our research project. I would also like to acknowledge the R esearch Technology Support Facility and the High Performance Computing C enter at Michigan State Univer sity for seque ncing and computational support as well as Channah Rock and Kurt Nolte (The University of Arizona) for helping with the Yuma field samplin g. Finally, I would like to thank my family for all their love and support. v TABLE OF CONTENTS LIST OF TABLES ................................ ................................ ................................ ...................... vii LIST OF FIGURES ................................ ................................ ................................ ..................... ix KEY TO ABBREVIATIONS ................................ ................................ ................................ ...... xi I. LITERATURE REVIEW ................................ ................................ ................................ ......... 1 1. Burden of Viral Foodborne Illness Associated with Fresh Produce on Public Health ...... 1 1.1 Foodborne Disease in the United States ................................ ................................ ........... 1 1.2 Foodborne Enteric Vir uses and Public Health 1.3 Significance of Fresh Produce as a Food Commodity and Vehicle of Pathogen Transmission ................................ ................................ ................................ ........................... 7 2. Methods for Studying Viruses in the Environment ................................ .............................. 8 3. Viruses in the Environment and Pre - Harvest Sources of Fresh Pro duce Contamination ............................................................................................................................. . 12 3.1 Viral Types, Characteristics, and Role in Food Industry ................................ ................. 12 3.2 Irrigation Water as an Environmental and Pre - Harvest Source of Fresh Produce Contamination ................................ ................................ ................................ ........................ 15 4. Viral Metagenomics ................................ ................................ ................................ ................. 18 4.1 Viral Metagenomic Technology ................................ ................................ ....................... 18 4.2 Metagenomic Insights into the Virome and Food Safety ................................ ................. 22 5. Post - Harvest Leafy Green Processing and Viral Contamination ................................ ....... 25 5.1 Leafy Green Commercial Processing Practices ................................ ............................... 25 5.2 Post - Harvest Fresh Produce Contamination and Viral Survival on Foods ...................... 27 II. III. METAGENOMIC IDENTIFI CATION OF VIRUS COMM UNITIES ASSOCIATED WITH LETTUCE AND IR RIGATION WATER ................................ ................................ ... 32 1. Introduction ................................ ................................ ................................ ............................. 32 2. Research Goals and Objectives ................................ ................................ .............................. 35 3. Materials and Methods ................................ ................................ ................................ .......... 35 3.1 Virus Recovery from Lettuce ................................ ................................ .......................... 35 3.1.1 Bacterial Host and Bacteriophage Preparation ................................ ....................... 35 3.1.2 Lettuce Inoculation, Elution, and Plating of Bacteriophag e ................................ ... 37 3.1.3 Plaque Concentration Assay and Percent Recovery Calculation ............................ 39 3.2 Virus Communities from Lettuce and Irrigation Water ................................ .................. 40 3.2.1 Irrigation Water and Lettuce Sample Collection ................................ ..................... 40 3.2.2 Virus Concentration and Purification ................................ ................................ ...... 42 3.2.3 Final Concent ration, Purification, and Nuc leic Acid Extraction ............................. 43 3.2.4 Random Amplification ................................ ................................ ............................. 43 3.2.5 Sequencing and Bioinformatics ................................ ................................ .............. 45 vi 4. Results ................................ ................................ ................................ ................................ .... 46 4.1 Virus Recovery Efficiency from Lettuce ................................ ................................ ........ 46 4.2 Yuma Irrigation Water Q uality ................................ ................................ ....................... 50 4.3 Irrigation Water Metagenomic S tatistics ................................ ................................ ........ 52 4.4 Irrigation Water V irome ................................ ................................ ................................ . 52 4.5 Yuma Lettuce Virome Statistics ................................ ................................ ....................... 69 4.6 Yuma Lettuce V irome ................................ ................................ ................................ ...... 71 5. Discussion ................................ ................................ ................................ ................................ .. 85 IV. VIRAL CROSS - CONTAMINATION OF LET TUCE DURING SMALL - SCALE LEAFY GREEN PROCESSI NG ................................ ................................ ............................... 92 1. Introduciton ................................ ................................ ................................ ............................. 92 2. Reseach Goals and Obje ctives ................................ ................................ ............................... 94 3. Materials and Methods ................................ ................................ ................................ ........... 95 3.1 Bacteriophage Inactivation and Free Chlorine Demand ................................ ................. 95 3.1.1 MS2 Inactivation during San i tizer Exposure ................................ ........................... 95 3.1.2 Coliphage MS2 Chlorine Demand ................................ ................................ ........... 96 3.2 Bacteriophage Reduction during Small - Scale Leafy Green Processin g With and Without a Chlorine - Based Sanitizer ................................ ................................ ................................ .... 96 3.2.1 Lettuce and Processing Line Preparation ................................ ................................ . 9 6 3.2.2 Lettuce Inoculation and Sampling ................................ ................................ ........... 97 3.2.3 Lettuce Processing and Plating ................................ ................................ ................ 98 3.2.4 Calculations and Statistical Analysis ................................ ................................ ....... 99 4. Results ................................ ................................ ................................ ................................ ... 100 4.1 MS2 I na ctiva tion during San i tizer E xposure ................................ ................................ . 100 4.2 Coliphage M S2 Chlorine D emand ................................ ................................ ................. 101 4.3 Virus Reduction during Leafy Green P rocessing ................................ ........................... 102 5. D 108 V. CONCLUSIONS, LIMITAT IONS, AND FUTURE NEE DS ................................ ........... 112 APPENDIX ................................ ................................ ................................ ................................ 115 REFERENCE S ................................ ................................ ................................ ........................... 136 vii LIST OF TABLES Table 1. Viral foodborne outbreaks and outbreak associated illnesses, hospitalizations, and deaths in the United Stat e s from 1998 - 2013 ................................ ................................ .................... 5 Table 2. Current methods used in virus detection ................................ ................................ ............ 9 Table 3 . Viral host examples and th eir significance in the environment and food industry .......... 14 Table 4. Advantages and disadvantages of current metagenomic next - generation sequencing technologies and examples applied to the food industry ................................ ................................ 19 Table 5. Bioinformatic computational techniques and tools used for viral metagenomics ........... 21 Table 6. Raw data and calculated eluent concentrations, lettuce concentrations, and percent recovery for P22 from inoculated romaine lettuce ................................ ................................ ........ 48 Table 7. Raw data and calculated eluent concentrations, lettuce concentrat ions, and percent recovery for MS2 from inoculated romaine lettuce ................................ ................................ ....... 49 Table 8 . Yuma irrigation water sample location descriptions and conditions ............................... 50 Table 9. Yuma irrigation water enterococci levels ................................ ................................ ........ 51 Table 10. Yuma irrigation water total coliform and E. coli levels ................................ ................ 51 Table 11. Yuma irrigation water s equence s tatistics f ollowing t rimming and a ssembly .............. 52 Table 12. Distribution of contigs larger than 200 bp for Yuma irrigation water virome .............. 53 Table 13. Yuma irrigation water viral host distribution ................................ ................................ 58 Table 14. Irrigation water viral family distribution representing greater than 3.0% of the virome ................................ ................................ ................................ ................................ ............ 62 Table 15. Bacteriophage host distribution of Yuma irrigation water samples .............................. 63 Table 16. Viral families and species of interest identified in irrigation water ............................... 66 Table 17. Yuma lettuce sequence statistics following trimming and assembly ............................ 70 Table 18. Distribution of contigs larger than 200 bp for iceberg and romaine lettuce virome ...... 72 Table 19. Iceberg and romaine lettuce viral host distribution ................................ ....................... 78 viii Table 20. Lettu ce viral family distribution representing greater than 3.0% of the virome ............ 82 Table 21. Yuma lettuce enteric virus nucleotide BLAST results ................................ .................. 84 Table 22 . Estimated MS2 concentration in inoculat ion suspension (PFU/mL) and on romaine lettuce following inocula tion (prior to centrifugal drying) ................................ .......................... 104 Table 23. Average MS2 concentration on romaine lettuce for triplicate lettuce samples collected following various stages of small - scale leafy greens processing ................................ ................. 104 Table 24. Log reduction of MS2 on romaine lettuce bet ween consecutive processing stages .... 107 Table 25. Flume water MS2 levels and reduction following processing ................................ ..... 108 Table A1. Raw contig sequence data used for Yuma irrigation water virome analysis .............. 120 Table A2. Raw contig sequence data used for Yuma lettuce vi rome analysis ............................ 124 Table A3. Raw data and the calculated MS2 eluent plaque concentration, lettuce concentration, and average lettuce concentration for samples collected during trial 1 leafy green processing without sanitizer ................................ ................................ ................................ ........................... 129 Table A4. Raw data and the c alculated MS2 eluent plaque concentration, lettuce concentration, and average lettuce concentration for samples collected during trial 2 leafy green processing without sanitizer ................................ ................................ ................................ ........................... 130 Table A5. Raw data and the calculated MS2 eluent plaque concentration, let tuce concentration, and average lettuce concentration for samples collected during trial 3 leafy green processing without sanitizer ................................ ................................ ................................ ........................... 131 Table A6. Raw data and the calculated MS2 eluent plaque concentration, lettuce concentration, and average lettuce concen tration for samples collected during trial 1 leafy green processing with 25 ppm free chlorine ................................ ................................ ................................ .................... 132 Table A7. Raw data and the calculated MS2 eluent plaque concentration, lettuce concentration, and average lettuce concentration for samples collected during t rial 2 leafy green processing with 25 ppm free chlorine ................................ ................................ ................................ .................... 133 Table A8. Raw data and the calculated MS2 eluent plaque concentration, lettuce concentration, and average lettuce concentration for samples collected during trial 3 leafy green processing with 25 ppm free chlorine ................................ ................................ ................................ .................... 134 Table A9. Flume water raw data and MS2 concentration for sanitizer experiments ................. 135 ix LIST OF FIGURES Figure 1. Distribution of contigs l arger than 200bp for Yuma irrigation water 1 virome ............. 54 Figure 2. Distribution of contigs lar ger than 200bp for Yuma irrigation water 2 virome ............. 54 Figure 3. Distribution of contigs l arger than 200bp for Yuma irrigation water 3 virome ............. 55 Figure 4. Distribution of contigs l arger than 200bp for Yuma irrigation water 4 virome ............. 55 Figure 5. Distribution of contigs l arger than 200bp for Yuma irrigation water 5 virome ............. 56 Figure 6. Distribution of contigs larger tha n 200bp for Yuma irrigation water 6 virome ............. 56 Figure 7. Viral host distribution in six Yuma irrigation water samples ................................ ......... 59 Figure 8. Yuma irrigatio n water viral family distribution ................................ ............................ 61 Figure 9. Irrigation water viral family distribution representing greater than 3.0% of the virome ................................ ................................ ................................ ................................ ............ 62 Figure 10. Iceberg lettuce - control genome distribution ................................ ............................... 73 Figure 11. Iceberg lettuce - worker harvest genome dist ribution ................................ .................. 73 Figure 12. Iceberg lettuce - post worker break genome distribution ................................ ............. 74 Figure 13. Romaine lettuce - control genome distribution ................................ ............................ 74 Figure 14. Romaine lettuce - worker harvest genome distribution ................................ ................ 75 Figure 15. Romaine lettuce - chop and wash geno me distribution ................................ ................ 75 Figure 16. Romaine lettuce - mixed salad genome distribution ................................ .................... 76 Figure 17. Yuma lettuce viral host distribution ................................ ................................ ............. 79 Figure 18. Yuma lettuce viral family distribution ................................ ................................ ......... 81 Figure 19. Lettuce viral family distribution representing greater th an 3.0% of the virome .......... 83 F igure 20. Coliphage MS2 inactivation after exposure to 25 ppm free chlorine in liquid suspension ................................ ................................ ................................ ................................ .... 101 x Figure 21. MS2 free chlorine demand ................................ ................................ ......................... 102 Figure 22 . MS2 reduction on inoculated romaine lettuce during small - scale leafy green p rocessing without sanitizer ................................ ................................ ................................ ......... 106 Figure 23. MS2 reduction on inoculated romaine lettuce during small - scale leafy green processing with sanitizer ................................ ................................ ................................ .............. 106 Figure A1. Yuma irrigation water sampling locations ................................ ................................ 116 Figure A2. Examples of Yuma furrow irrigation wa ter (1), iceberg lettuce following harvest (2), and romaine lettuce following harvest (3) ................................ ................................ 117 Figure A3. Ultrafiltration system for the primary concentration of irrigation water samples .... 118 Figure A4. Lettuce centrifugal dryer (1), shredder (2), flume tank (3), and mechanical shaker table (4) sampling locations ................................ ................................ ................................ ......... 119 xi KEY TO ABBREVIATIONS BLAST Basic Local Alignment Search Tool bp Base Pair CDC Centers for Disease Control and Prevention CFSAN Center for Food Safety and Applied Nutritio n Contig Contiguous Sequence DNA Deoxyribonucleic Acid ds Double - Stranded eFORS E lectronic Foodborne Outbreak Reporting System ELISA Enzyme - Linked Immunosorbant Assay E - Value Expected Value FDA U.S. Food and Drug Administration FDOSS Foodborne D isease Outbreak Surveillance System FOOD Foodborne Outbreak Online Database FoodNet Foodborne Diseases Active Surveillance Network HACCP Hazard Analysis and Critical Control Points HPAI Highly Pathogenic Avian Influenza Virus IDBA - UD Iterative de Bru ijn Graph Assembler IFIC International Food Information Foundation Council kb Kilobase MEGAN Metagenome Analyzer MMWR Morbidity and Mortality Weekly Report MPN Most Probable Number xii NCBI National Center for Biotechnology Informa tion NFPA National Food Processors Association NGS Next - Generation Sequencing NORS National Outbreak Reporting System PBS Phosphate Buffered Saline PBW Phosphate Buffered Water PCR Polymerase Chain Reaction PEG Polyethylene Glycol PFU Plaque F orming Unit qPCR Quantitative PCR RefSeq Viral Reference Sequence RTE Ready - to - Eat RT - PCR Reverse Transcriptase PCR RNA Ribonucleic Acid ss Single Stranded STEC Shiga toxin - producing Escherichia coli TSA Trypticase Soy Agar TSB Tryptic S oy Broth USDA United States Department of Agriculture WHO World Health Organizatio n 1 I. LITERATURE REVIEW 1. Burden of Viral Foodborne Illness Associated with Fresh Produce on Public Health 1. 1 Foodborne D isease in the United States Foodborne diseas e is a serious threat to public health and food safety worldwide. The World Health Organization (WHO) describes foodborne illness as iseases, usually either infectious or toxic in nature, caused by agents that enter the body through the ingestion of foo d (CDC , 2013 a ) . C urrently WHO estimates that d iarrheal diseases many of which are attributed to contaminated food resu lt in 2 million deaths annually ( CDC , 2013 a ) . In the United States, f ood safety monitoring efforts in the food supply chain have greatl y increased (Crutchfield & Roberts, 2000) . With the implementation of Hazard Anal ysis and Critical Control Point (HACCP) programs and food safety initiatives including the National Food Safety Initiative of 1997 and the Food and Dru g Administration (FDA) Food Safety Modernization Act of 2011 (Crutchfield & Roberts, 2000, Oliver et al. , 2009) . However, d espite recent advances in food monitoring, f oodborn e outbreaks remain a serious threat to public health and it has been suggested that consumer confidence in food safety has slowly begun to decline (IFIC, 2014) . According to the Centers for Disease Control and Prevention (CDC), approximately 48 million pe ople in the United States are expected to be affected by foodborne illnesses each year ( CDC, 2014 ). Of the 48 million people affected, approximately 9.4 million (20%) cases of illness are caused by 31 known foodborne pathogens while a staggering 38.4 mill ion (80%) are caused by unspecified agents ( CDC , 2014 ). This remarkably high portion of unspecified agents suggests a current sho rtage of data collection from affected patients at local health agencies and ultimately a current knowledge gap concerning unk nown or unidentified foodborne agents. 2 A foodborne outbreak occurs when a group of individuals consume contaminated food and two or more of them develop the same illness. In 2013 alone , an estimated 8 1 8 foodborne dise ase outbreaks resulted in 13,360 ca ses of illness, 1,062 hospitalizations, and 16 deaths in the United States ( CDC, 2015 b ) . Consequently, foodborne illness is estimated to be costly for the health care system. The approximate annual cost due to foodborne illness from pathogenic bacteria, parasites, and viruses combined in the U.S. ranges from $51.0 to $77.7 billion , varying with the U.S. Dep artment of Agriculture (USDA) and U.S. Food and Drug A dministration (FDA) cost - of - illness model s utilized (Scharff, 2012) . Currently there are a la rge variety of bacterial, viral, and parasitic human pathogens associa ted with foodborne disease. The bacterial pathogens causing the highest number of repo rted foodborne outbreaks in 2013 were Salmonella spp. , Shiga toxin - producing Escherichia coli (STE C), Clostridium perfringe n s, Campylobacter spp. , and Vibrio parahaemolyticus accordingly ( CDC, 2015 b ). Furthermore, viral foodborne pathogens were dominated by norovirus while Cryptosporidium parvum , G iardia lambia and T richinella spiralis were the most c ommonly reported parasitic infections ( CDC, 2015 b ) . Despite recognition of these human pathogens in our food supply, foodborne infections conti nue to emerge and in some cases have risen in recent years (CDC, 2011 ) . Viruses in pa rticular are becoming incr easingly recognized as foodborne pathogens, with an increasing number of outbreaks occurring between 1998 and 2008 (Gould et al., 2013) . To date, numerous foodborne enteric viruses causing gastroenteritis in humans have been identified , however , there are still a variety of human viruses that are capable of replication within the intestinal tract and their role in our food sys tems is currently unknown. These data indicate the v ulnerability of our food system to contamination and emphasize the 3 need to iden tify current knowledge gaps in food safety , particularly in relation to foodborne viruses . 1.2 Foodborne E nteric Viruses and Public H ealth Human enteric viruses, which are commonly transmitted through the fecal - oral route , are a serious threat to public h ealth and safety . Enteric viruses have high infectivity rates (10 - 100 particles result in high probabilit y of infection) and many lack an envelope which allows for resistance to thermal inactivation , facilit ating virus survival and maintaining infectivity in the environment (Fong & Lipp, 2 005 , Gibson & Schwab, 2011, Newell et al., 2010) . These viruses concentrations (10 5 - 10 11 viral particles per gram of stool) in to the environment through the feces or vomit of an infected individual (Fong & Lipp, 2005 ) . Water and food can then become contaminated at both pre - and post - harvest levels of production . Viruses are increasingly being recognized as water and foodborn e etiological agents. There are hundreds of human pathogenic water and foodborne viruses of fecal origin including adenoviruses, astrovirus, norovirus (genotypes I and II) , polioviruses, enteroviruses, hepatitis A and E virus, sapoviruses, reoviruses , and rotaviruses which can cause gastroenteritis, meningitis, liver disease, infantile diarrhea, respiratory illness, or neurological symptoms (Cook , 2013, Fong and Lipp , 2005, Koopmans et al. 2004). However, many symptoms associated with these human pathogen s are generally mild or self - limiting and therefore many infections are underreported (Fong & Lipp, 2005, , Seymour & Appleton , 2001 ) . More severe cases that result in illness, hospitalization, and death are usually observed in children, the elderly, and immunocompromised individuals (Fo ng & Lipp, 2005, Rodríguez - Lázaro et al., 2012) . 4 To date, the viruses most frequently associated with foodborne illness include human norovirus and hepatitis A virus (Koopmans & Duizer, 2004, Newell et al., 2010) . Norovirus is a non - enveloped, non - culturable, single - stranded RNA virus in the Calciviridae family and is the leading cause of viral gastroenteritis worldwide (DiCaprio et al. 2004). This enteric virus has a short incubation period (12 - 48 hours) a nd is commonly associated with outbreaks on cruise ships, however, it is often difficult to confirm the specific food and water source of transmission (Isakbaeva et al., 2005, Kroneman et al., 2008) . Hepatitis A is a non - enveloped, ssRNA virus of the Pico rnaviridae family that exhibits slow replication in culture and has a significantly longer incubation period (15 - the sixth leading cause of foodborne disease in the United States, however introduction of the hepatitis A viral vaccine in 1995 led to a significant drop in hepatitis A infections (Seymour & Appleton, 2001) . In the United States, the CDC Morbidity and Mortality Weekly R eport (MMWR) named viruses as the primary cause o f foodborne outbreaks with known etiology in a survey of foodborne outbreaks from 1998 to 2008 (Gould et al., 2013) . The CDC monitors and gathers data on foodborne outbreaks in the United States through the Foodborne Disease Outbreak Surveillance System (FDOSS) , which relies on public health agency reporting through the electronic Foodborne Outbreak Reporting System (eFORS). In 2009, eFORS was replaced by the National Outbreak Reporting System (NORS) and data on foodborne outbreaks from 1998 - 2013 can now be viewed online through the Foodborne Outbreak Online Database (FOOD). Table 1 shows the updated viral foodborne outbreak data from 199 8 - 2013. Data from 1998 - 2008 were obtained from Gould et al. ( 2013 ) and 2009 - 2013 data were added using viral o utbreak data obtained in FOOD. 5 * Data obtained from the CDC Foodborne Outbreak Online Database **Definitions: CE = confirmed etiology; SE = suspected etiology. Etiologies are confirmed using laboratory and clinical guidelines ( CDC, 2015a ). Those that do not meet guidelines are labeled suspected etiology. ***Re produced from Gould et al. 2013 Table 1. Viral foodborne outbreaks and outbreak associated illnesses, h ospitaliza tions, and d eaths in the United Stat e s from 1998 - 2013* Outbreaks Illnesses Hospitalizations Deaths Total Total Total Total Etiology ** CE SE # % CE SE # % CE SE # % CE SE # % Astrovirus 1 1 2 0 14 22 36 0 0 0 0 0 0 0 0 0 Hepatitis A 85 1 86 2 2370 4 2374 2 363 0 363 22 8 0 8 35 Norovirus 2,786 1,936 4,722 96 92,339 34,629 126,968 96 967 300 1,267 76 7 0 7 30 Rotavirus 4 8 12 0 204 110 314 0 0 5 5 0 7 1 8 35 Other viral 8 96 104 2 510 2,568 3,078 2 7 18 25 2 0 0 0 0 Total 2,884 2,042 4,926 100 95,437 37,333 132,770 100 1,337 323 1,660 100 22 1 23 100 6 From 1998 - 2013 food borne viruses caused 4,926 outbreaks, 132,770 cases of illness, 1,660 hospitalizations and 23 deaths. Norovirus was the most frequently reported foodborne viral pathogen, accounting for 96% of viral outbreaks, 96% of illnesses and 76% of hospitalizations. Hepatitis A (86), rotavirus (12), and astrovirus (2) outbreaks were also reported. Interestingly, r ecent foodborne outbreak s also suggest that viruses are commonly associated with fresh produce contamination and transmission. Studies on norovirus in pa rticular have shown strong associations with fresh produce consumption (Dicap rio et al. , 2012, Gould et al., 2013, Widdowson et al., 2005) . It has been suggested that norovirus is responsible for over 40% of the annual fresh produce outbreaks in the Unit ed States (DiCaprio et al. 2012, Seymour & Appleton, 2001). Furthermore, CDC data from 1998 - 2008 labeled norovirus and leafy greens as the pathogen - commodity pair most likely to be associated with a foodborne outbreak, causing a total of 4,011 illnesses ( Gould et al. 2013). Hepatitis A has also been linked to multiple outbreaks in green onion, blueberries, and strawberries (Calder et al., 2003, Hutin et al., 1999, Wheeler et al., 2005) . Although norovirus and hepatitis A have been frequently associated w ith fresh produce outbreaks, there is currently limited knowledge on the role of other viruses in our food systems. Despite increased awareness of viruses as foodborne disease agents, the Foodborne Diseases Active Surveillance Network (FoodNet) which act ively monitors trends in f oodborne illnesses and assesses food safety initiative impacts in the U.S. continues to only monitor bacteria ( Campylobacter, Listeria , Salmonella , STEC O157 and non - O157 , Shigella , Vibrio , and Yersinia ) and parasites ( Cryptosporidium , Cyclospora ) . However in addition to reporting through FDOSS, the CDC has begun to actively monitor viral outbr eaks, namely norovirus, through the CaliciNet (2009) and NoroSTAT (2012) surveillance systems (CDC, 2013b). 7 Although these surveillance systems have led to increased awareness of the role of viruses in the edge on where contamination occurs in the supply chain and other viral pathogens of concern. 1.3 Significance of Fresh Produce as a Food Commodity and Vehicle of Pathogen Transmission Fresh produce is a food commodity of increasing public health interest . Fruits and vegetables compose two of the five basic food groups recommended by the USDA as components of a healthy diet. Studies on food consumption trends in the United States have shown increased per capita consumption of fruits and vegetables. Acco rding to an agricultural economist with the USDA, average annual per capita fruit and vegetable consumption per pound increased by 25% between 1977 - 1979 and 1997 - 1999 (Pollack, 2001) . In another study, per capita consumption of fruit and vegetables in the United States was estimated to increase 19% and 29% respectively between 1980 and 2001 (Clemens, 2004) . Suggested drivers of increased fresh produce consumption include increased production, product convenience, improved technology that maintains produce quality, greater availability, and consumer desire to maintain a healthy lifestyle (Pollack, 2001) . For example, from 1970 to 2012, the average amount of fresh vegetables and fruits available for consumption increased by 67 and 6 pounds, respectively (US DA, 2014). In response to the changing fresh produce supply and demand, traditional agricultural and post - harvest practices have been altered. Practices such as cutting and coring at harvest, increased importation and transportation, and large scale prod uction facilities are now employed to support changing consumer habits (Heaton et al. 2008, Lynch et al. 2009). 8 O utbreaks associated with fresh produce are becoming increasingly recognized. In addition to intensive prod uction and processing practices, fresh produce is commonly consumed raw, making it an ideal vehicle for pathogen transmission. An analysis of the FDOSS found an increasing number of foodborne outbreaks associated with the consumption of raw produce in the U.S. , rising from 0 .7% in 1970 to 6 (Sivapalasingam et al, 2004 ) . Of particular interest, the food items most commonly associated with fresh produce outbreaks were leafy greens such as lettuce and salads in addition to melon, sprouts, and berries (Sivapalasingam et al., 2004) . In a more recent stud y, Painter et al. ( 2013 ) used F D OSS to summarize data on foodborne outbreaks, illnesses, and hospitalizations attributed to 17 mutually exclusive food commodities. According to this study, from 1998 to 2008 an estimated 46% o f annual illnesses, 38% of annual hospitalizations, and 23% of annual deaths acquired in the United States were attributed to fresh produce ; w hereas meat and poultry contributed 22% and 29% of the illnesses and deaths (dairy and eggs; fish and shellfish co ntributed 20% and 15% an d 6.1% and 6.4%, respectively) (Painter et al., 2013) . Furthermore, this study found that leafy green vegetables were the food commodity responsible for the highest number of foo dborne illnesses (2.2 million) (Painter et al., 2013) . More recently, in 2013 leafy vegetables were implicated in 9 foodborne outbreaks which resulted in 207 cases of illness (CDC, 2015 b ). Despite increasing r ecognition of fresh produce as a vehicle for pathogen transmission, foodborne outbreaks attributed to fresh produce remain a public health and food safety concern. 2. Methods for Studying Viruses in the E nvironment A summary of the current viral detection methods in food and water is provided in Table 2 . Historically, cell culture has served as the method for viral d etection and 9 d iscovery . Bacteriophage were among the first viruses to be replicated in vitro due to easy lab oratory based manipulation of bacterial hosts. To date, the most common method of bacteriophage isolation, purification, and enumeration is the plaque assay, which dates back to its original discovery in 1917 , 1917, Kutter & Sul akvelidze, 2004) . In this method, bacterial host cells are exposed to a virus that upon infection lyses the surrounding cells, resulting in a clear zone in the agar medium called a plaque which represents a single infectious viral particle. Although this detection method has proven Table 2 . Current methods u sed in virus d etection Detection Met hod Method d escription Current e xamples Advantages Disadvantages Cell c ulture Viruses infect and replicate in host specific cells Continuous c ulture lines from animal cells Direct isolation of a variety of cultivable viruses to high titers Many virus es uncultivable, requires specific cell line, costly, time consuming In v itro bacterial host culture Viruses infect and lyse bacterial host cells Bacteriophage plaque assay Direct bacteriophage isolation, purification, and enumeration Requires specific bacterial host, issues in reproducibility (diluting, plaque size, incomplete lysis, pl aque aggregation) Electron microscopy Microscope that uses an electron beam to illuminate and magnify viruses in detail Transmission and scanning electron microscopes Does not require prior knowledge of organism DNA, provides high resolution image Poor detection limit, need high concentrations, maintenance, cumbersome, training, cannot identify virus ELISA Antigen - antibody pathogen detection Indirect and sandwich E LISA Quick and rapid detection, by - pass cell culture Require a specific probe for detection, not applicable for virus discovery PCR; quantitative PCR (qPCR) Viral DNA or RNA amplification and enumeration of a known sequence Reverse transcriptase PCR (RT - PCR ) , integrated cell culture PCR Fast, high throughput, quantitative data, sensitivity, repeatability Must know the sequence. Cost of equipment and reagents, reaction inhibition, data analysis, training. Next generation sequencing (metagenomics) High th roughput viral DNA or RNA sequencing all genomes in an environmental sample Pyrosequencing, sequencing - by - synthesis, sequencing - by - ligation Fast, high throughput, cost of sequencing, reliable identification of microbial communities Short sequence read le ngth, time and training required for bioinformatics analysis 10 beneficial in the field of virology, cell culture and in vitro virus replication has many disadvantages. A major limitation of vial studies in cell culture or in vitro is that this method only targets viruses capable of replic ati ng in cells that can be propagated , all of which require a specific cell line for virus proliferation. Such viruses include adenovirus, enteroviruses (poliovirus, coxsackie viruses, echoviruses), influenza A and B, Measles virus, Mumps virus, rhinoviru s, Ebola, SARS - coV, VZV and hMPV (Leland & Ginocchio, 2007) . Many other viruses, including those important in foodborne disease, cannot replicate in cell culture and therefore require a different method of detection. Electron microscopy is another trad itional method used in virus detection. Electron microscopes (scanning and transmission) use a beam of electrons to illuminate and magnify viruses in great detail. In 1939, the first virus (tobacco mosai c virus) was visualized using electron microscopy a nd this technology has since then aided in the discovery of viruses such as smallpox and poliovirus (Goldsmith & Miller, 2009). Although prior knowledge of organism DNA is not necessary for detection, disadvantages include the need for high viral concentr ations, poor detection limits, and the inability to identify the virus beyond the family level. To date, the serological method most commonly used in virus research is the Enzyme - Linked Immunosorbant Assay (ELISA). In this technique, a viral antigen imm obilized to a solid surface binds to a specific antibody which is either linked to an enzyme or can be detected by a secondary antibody linked to an enzyme. After adding an enzymatic substrate, a visible signal such as color change is produced and the ant igen can be quantified. This technology has been regularly applied in plant virus d etection (as early as 1976) as well as food authenticity in the food industry (Voller et al. 1976, Asensio, González et al. , 2008) . Although serological methods 11 bypass the need for cell culture and are time and cost effective, they require a specific probe for virus detection and are not applicable to virus discovery. To date advanced filtration and molecular detection methods have greatly improved virus detection and mo nitoring in environmental samples, especially for enteric viruses in water systems. Specifically tangential - flow, hollow fiber ultrafiltration allows for virus concentration based on size exclusion (molecular weight cutoff) from large volumes of water (Gi bson & Schwab, 2011, Liu et al., 2012, Smith & Hill, 2009) . This technology has been readily applied to concentrate viruses from water systems (reclaimed and surface) and when combined with molecular detection techniques, has provided a better understandi ng of the microbial quality of water (Gibson & Schwab, 2011, Liu et al., 2012) . Viruses are traditionally further concentrated by passing the remaining filtrate through a 0.22 µm filter (bacteria removal), polyethylene glycol (PEG) precipitation, and ultr acentrifugation prior to molecular detection (Croci et al., 2008, Liu et al., 2012, Rosario et al. , 2009) . The following sequence - and culture - dependent molecular methods have been used for direct and rapid detection of viruses in the environment 1) poly merase chain reaction (PCR), reverse transcription PCR (RT - PCR), and integrated cell culture PCR, 2) quantitative PCR (qPCR), 3) Sanger sequencing, 4) and whole genome sequencing. artlett & Stirling, 2003). Since its discovery, traditional endpoint PCR has been modified to better detect viruses in clinical and environmental samples. Many foodborne enteric viruses (norovirus, hepatitis A virus, astrovirus, rotavirus, enterovirus) a re composed of RNA and direct detection requires reverse transcription. To date, RT - PCR remains a gold standard for enteric virus detection, especially in foods (Bidawid et al., 2000, Hyeon et al., 2011, Leggitt & Jaykus, 2000, 12 Love et al. , 2008) . Integ rated cell culture PCR combines culture with molecular methods to detect viruses in the environment (Reynolds, 2004) . This technique has been frequently applied to study enteric viruses in multiple water types including drinking water, river water, and se wage, but has also been combined with qPCR technology to detect enteric viruses (Hepatitis A) in fresh produce (Greening et al., 2002, Hyeon et al., 2011, Lee & Jeong, 2004) . Unlike the conventional end - fluorescent technology to monitor and quantify targeted nucleic acids as they are amplified (Aw & Rose, 2012, Fraga 2014 ) . To date, numerous studies have used RT - qPCR to detect enteric viruses in water sources (Aw & Rose, 2012) . More recently, studies h ave begun to investigate methods of enteric virus recovery and RT - qPCR detection in fresh produce, including vegetables (lettuce, chicory, spinach, mixed salads, green onion, basil) and fruits (strawberries, raspberries, blueberries, blackberries) (Butot e t al., 2007, Dubois et al., 2007, Sánchez et al. , 2012) . In the past decade, the development of new genomic technologies has been exceedingly important for the discovery of novel viruses. More specifically, culture - and sequence - independent sequencing technologies, known as next - generation sequencing (NGS), allow for the examination of entire microbial communities in an environmental sample (metagenomics) and do not require previous knowledge of viral nucleic acid sequences. These new technologies and metagenomic techniques are now being used to s tudy viruses in the environment to gain insights into the virus world. 3. Viruses in the Envir onment and Pre - Harvest Sources o f Fresh Produce Contamination 3.1 Viral Types, Characteristics, and Role in Food I ndustry Viruses are intracellular, infectious agents that are ubiquitous in nature and replicate within the cells of living organisms to cause a wide range of diseases. The size, shape, chemical 13 structure, and genome composition are all characteristics used to classify viruses. In the Baltimore classification system, viruses are grouped into families based on mode of replication and genome composition, which includes double - stranded (ds) and single - stranded (ss) DNA and RNA. In addition, viruses are h ost specific and are often identified by the hosts they infect. Examples of viral hosts include bacteria (bacteriophage), algae, plants, animals (invertebrate and vertebrate), and humans. The significance of these viral hosts in the environment and popul ar examples of each virus type are included in Table 3. Viruses that infect bacteria, called bacteriophage or phage for short, are diverse and widely distributed in the environment (Breitbart & Rohwer, 2005) . Bacteriophage are of interest to the food in dustry because they provide insight into the bacterial host populations in the environment and the host specificity of viruses can be used as potential indicators of fecal contamination (Aw et al. , 2014 , Leclerc et al., 2000) . In addition, male - specific, non - enveloped, RNA bacteriophage such as MS2 are shown to have similar resistance and survival characteristics as enteric viruses (including norovirus) in water and fresh produce and can therefore act as a surrogate for foodborne viruses that cannot be pro pagated in cell culture (Dawson , 2005, Havelaar et al. , 1993) Furthermore, studies have shown that specific lytic bacteriophage are a promising tool for reducing bacterial pathogens on fresh produce (Sharma, 2013) . Viruses that infect algae, which are a quatic chlorophyll containing unicellular and multicellular organisms, are diverse and prevalent in aquatic ecosystems. These viruses can act as mortality agents to control algal host populations and are possible constituents of irrigation water (varying with source). Pathogenic viruses that infect and cause disease in staple crops are of primary concern to the food industry. Plant viruses that cause physical and chemical alterations to fruits and vegetables are responsible for major losses in crop produ ctivity, yield, 14 quality, and ultimately are costly to the fresh produce industry. Invertebrates (insects, Table 3. Viral host examples and their significance in the environment and food industry Host Significance Viral e xamples Reference(s) Bacteria (b acteriophage) Control pathogenic bacteria in the food chain Biogeochemical cycling Antibiotic resistance Potential indicators of fecal contamination Somatic Phage: T even phages ( Myoviridae ( Siphoviridae ), P22 ( Podoviridae ), phi X174 ( Microviridae ) F + specific phage: MS2 ( Leviviridae ), CTX ( Inoviridae ) ( Ackermann, 2009, Colomer - Lluch et al., 2011, Davis et al., 2000, Leclerc et al. , 2000, Rodríguez - Lázaro et al., 2012) Algae Regu late fresh water and marine food webs Biogeochemical cycling Assist in algal bloom reduction and formation dsDNA viruses of the Phycodnaviridae family ( Chlorovirus , Coccolithovirus , Prasinovirus , Prymnesiovirus , Phaeovirus and Raphidovirus) (Baudoux & B russaard, 2005, Brussaard, 2004, Wilson et al. , 2006) Plants Plant disease: leaves and fruit spotting, ringspots, discoloration, reduced vegetative output, poor growth High economic costs for fresh produce food industry due to reduced crop quality, produ ctivity, and yield Mosaic viruses: Cucumber, tobacco, tomato ( Cucumovirus, Tobamovirus ) Tomato spotted wilt virus ( Tospovirus ) Potato virus X and Y ( Potexvirus, Potyvirus ) Plum pox potyvirus ( Potyvirus ) (Mehle & Ravnikar, 2012, Rybicki, 2015) Invertebra te a nimals Vector borne virus transmission to animals and humans Biological control agents for management of agroecosystems, stored products, and forestry High economic costs for food industry and aquaculture (loss of productivity) Vector borne vi ruses: Yellow fever, Dengue fever, West Nile, Japanese encephalitis ( Flavivirus ) Viruses as biological control agents: dsDNA Baculoviruses Pathogenic marine invertebrate viruses: Baculoviruses, Iridoviruses, Reoviruses, Rhabdoviruses (Johnson, 1984, Lac ey et al., 2001) Vertebrate a nimals Viral zoonotic illnesses High economic costs due to livestock productivity loss Livestock pathogens: Foot - and - mouth disease, bovine viral diarrhea, Newcastle disease Emerging viruses: Swine hepatitis E, Nipah virus, SA RS Coronavirus , highly pathog enic avian influenza virus (HPAI - H5N1) (Chi et al. , 2002, FAO/WHO, 2008, Pimentel et al., 2001) 15 Table 3 ( ) Humans Threat to public health and food safety High economic costs to public health system Food and water born e viruses: Norovirus and Sapovirus ( Caliciviridae ) , enterovirus and hepatitis A ( Picornaviridae ), adenovirus ( Adenoviridiae ), and astrovirus ( Astroviridae ) (Fong & Lipp, 2005, Seymour & Appleton) species and therefore are infected by a wide variety of vir uses, many of which also infect mammal, bird, and plant species. Invertebrate viruses may directly or indirectly impact numerous food commodities by infecting and causing disease in agricultural pests as well as food fish and shrimp species. For vertebra te animals, viruses infecting livestock that ultimately impact food production and safety are of primary importance. Although many animal viruses that cause disease in humans (zoonotic) are transmitted by direct contact rather than through a food vehicle, emerging viruses such as swine hepatitis E, Nipah virus, SARS Coronavirus and highly pathogenic avian influenza virus (HPAI - H5N1) are currently suspected of foodborne transmission (FAO/WHO 2008) . This shows the increasing importance of animal viruses in our food systems and public health. 3.2. Irrigation Water as an Environmental and Pre - Harvest Source of Fresh Produce Contamination Fresh produce, which is subject to intense production practices, has many opportunities for human pathogen contamination from farm - to - fork. The HACCP principles and guidelines were established in 1997 to help guide the food industry in identifying, evaluating, and ultimately controlling chemical, biological, and physical hazards in food s throughout the food supply chain (FD A, 2014). Recognizing that sources of contamination vary with hazard type, food commodity, and stage of production (pre - vs. post - harvest), the HACCP system provides the food industry with recommendations on how to control possible hazards and critical co ntrol points based on individual practices. For the fresh produce industry at the pre - harvest level, 16 irrigation water and runoff, soil, fertilizer, animals, insects, and food handlers have all been implicated as sources of human pathogen contamination in fresh produce (Heaton & Jones, 2008, Lynch et al. , 2009) . The microbial quality of irrigation water is now recognized as one of the primary pre - harvest factors influencing the microbial quality of fresh produce. In 2006, two separate outbreaks of E.coli O157:H7 were attributed to leafy greens (spinach and iceberg lettuce) grown in California. An investigation of the watersheds surrounding the farms suggested that water used to irrigate the crops was the likely source of contamination (Gelting & Baloch, 2012) . Regarding viruses, sewage contaminated irrigation water has previously been implicated as the source of Hepatitis A outbreaks attributed to fresh produce, however in many cases this could not be proven (Seymour & Appleton, 2001) . Surveillance and on - sight field investigations of the environmental sources of fresh produce contamination, such as irrigation water, during foodborne viral outbreaks is limited by current methods for virus detection. Microbial quality standards for irrigation water, wh en they exist, vary greatly between countries (and states) as well as by water source. Compared to groundwater, reclaimed and surface irrigation water sources are more susceptible to human pathogen contamination and therefore may have recommended guidelin es for agricultural use (Steele & Odumeru, 2004) . Common microbial indicators used in irrigation water quality guidelines include coliform bacteria (total and fecal), E. coli , enterococci, and nematode eggs (Steele & Odumeru, 2004 , EPA, 2012 ) . For reclai med water used in the irrigation of food crops intended for human consumption, the EPA currently recommends daily monitoring of fecal coliforms, with no detectable fecal coliforms per 100 mL of water (EPA , 2012). For surface waters, the EPA recommends few er than 1000 fecal coliforms per 100 mL of irrigation water for use on crops 17 ( Steele & Odumeru, 2004 ). However studies have shown that viral pathogens do not correlate with traditional indicators in water and the application of enteric viruses or coliphag e as indicators of fecal contamination has not yet been utilized for monitoring irrigation water quality (Harwood et al., 2005) . A few studies have investigated the relationship between foodborne bacterial pathogens present in irrigation water and recip ient fresh produce (Heaton & Jones, 2008) . For viruses, studies that focus on detecting enteric viruses in irrigation water or in foods exist, however few studies have investigated the relationship between viruses in irrigation water and recipient produce . In one hydroponic study, DiCaprio et al. (2012) found that human norovirus and animal caliciviruses (Tulane virus and murine norovirus) were capable of efficient internalization and dissemination in romaine lettuce when introduced into the feed water (D icaprio et al., 2012) . Here internalization refers to virus entry into the plant interior tissues (not virus infection of plant cells). This suggests that viruses present in irrigation water can also occupy plant tissues of recipient crops and ultimately impact their viral composition. A study by Stine et al. (2005) determined the concentration of hepatitis A in irrigation water needed for a 1:10,000 yearly risk of infection from consuming irrigated fresh produce to ultimately help guide microbial stan dards in irrigation water. In this study field - grown cantaloupe, iceberg lettuce, and bell peppers were drip (target plant roots) or furrow (flooded channels) irrigated with coliphage PRD1 (surrogate for Hepatitis A) and a quantitative microbial risk asse ssment was conducted. Results suggest that risk of infection varies with crop type, irrigation method, and time between irrigation and harvest for consumption. Specifically, direct targeting of the plant roots using subsurface drip irrigation was found t o reduce the risk of crop contamination (Stine et al., 2005) . The use of subsurface drip irrigation as a way to mitigate 1 8 fresh produce contamination of viral pathogens has been supported in other studies (Alum et al., 2011, Song et al. , 2006) . However in a study by Choi et al. (2004) where MS2 and PRD1 phage inoculated irrigation water was used to cultivate field lettuce, lettuce virus levels were higher using the subsurface drip method when compared to the furrow method . This study suggested direct cont act of irrigation water with lettuce stems in addition to shallow drip irrigation depth as likely causes of increased contamination (Choi et al., 2004) . More recently, a study by Cheong et al. (2009) used reverse transcription and cell culture PCR to de tect norovirus, enteroviruses, adenoviruses, and rotaviruses in surface applied irrigation groundwater and recipient fresh produce (cherry tomato, chicory, cabbage, beet, lettuce, spinach) . Although a clear relationship between enteric viruses in irrigati on water and recipient fresh produce was not detected, this study found that 1) virus occurrence did not relate to coliform (total and fecal) or enterococci levels traditionally used to assess microbial quality and 2) irrigation water and fresh produce sam ples positive for enteric viruses were collected during the same time period (irrespective of sampling location). These results suggest that bacterial indictors do not accurately represent all microbial hazards in water and enteric virus presence in irrig ation water and fresh produce may vary seasonally. The relationship between irrigation water and fresh produce viral contamination could be better understood if greater assessments could be undertaken to provide more resolution on the types of viruses pre sent. This is now possible using novel metagenomics techniques which stud y entire microbial communities. 4. Viral Metagenomics 4.1 Viral Metagenomic Technology Metagenomics is defined as the study of all genetic material from a mixed community of organ isms (Handelsman, 2004) . This approach is a sequence - and culture - independent method 19 for studying entire microbial communities in an environmental sample. In 1998, the first commercially available high - throughput instruments the GE Healthcare MegaBACE 10 00 and ABI Prism 3700 DNA Analyzer, used a combination of traditional Sanger and capillary sequencing to crea te these large DNA datasets (Kircher & Kelso, 2010) . However, i n the past decade, innovative sequencing systems, including Roche 454 G enome S equen cers (Junior, Junior+, and FLX Titanium), Illumina (Genome analyzer (I, IIx, IIe) , Miseq, and Hiseq), and the Applied Biosystems SOLiD sequencing platforms , ha ve resulted in greater daily throughput and significant c ost - reductions (Kircher & Kelso, 2010, L iu et al., 2012) . These NGS technologies are the current metagenomic tools most commonly used to identify microbial communities from a wide variety of environmental samples and are compared in Table 4. Table 4. Advantages and disadvantages of current me tagenomic next - generation sequencing technologies * and exampl es applied to the food industry Sequencing t echnology Sequencing m ethod ** Advantages Disadvantages Current metagenomic research (f ood industry a pplication) References Roche 454 Genome Sequence r (junior and FLX systems) Pyrosequencing Larger read length Low er throughput, higher error rate, high reagent costs Food microbiota (cheese, fresh produce, fermented foods); changes in microbiota during food processing, fermentation, and storage; microbio ta in irrigation water (Ercolini, 2013, Leff & Fierer, 2013, Lopez - Velasco et al. , 2011, Ottesen et al., 2013, Park et al., 2011) Illumina (Genome Analyzer, Hiseq, Miseq) Sequencing - by - synthesis High throughput, lower sequencing cost Short read length Al cohol fermentation, rumen microbiome and virome in cattle (Ercolini, 2013, Ross et al., 2012, Ross et al. , 2013) Applied Biosystems SOLiD Sequencing - by - ligation Increased accuracy Short read length None *** *R eproduced from Liu et al. 2012 ** Definition s: Pyrosequencing is a method based on the detection of pyrophosphates released during DNA polymerase synthesis (nucleotide incorporation). Illumina sequencing - by - synthesis relies on the detection of single bases (base - by - base sequencing) as they are inco rporated into DNA strands by DNA polymerase. Sequencing - by - ligation uses a DNA ligase enzyme to identify fluorescently labeled oligonucleotide probes and perform the sequencing reaction. ** * Could not find examples of this technology applied to foods or fo od systems 20 The Roche 454 and Illumina systems are the NGS technologies most frequently applied to viral metagenomics (Mokili et al. , 2012) . The Roche 454 GS FLX system has been called the nior system w hich fits on a benchtop, the FLX is a large system for accurate, high - throughput sequencing that results in long read lengths of DNA (up to 1 kb) ( Roche , 2015 ). This system is well suited for large genomic projects and can be used for pathogen detection i n complex environmental samples. To date, the Genome Analyzer (IIx), Miseq , NextSeq, Hiseq, and HiSeqx all use the latest Illumina sequencing technologies, with Illumina Hiseq providing the necessary sequencing power for studying la rge scale production ge nomics. Using base - by - base sequencing, these NGS systems provide more information tha n traditional Sanger sequencing and can be used to sequence whole genomes, target regions, RNA, and entire microbial communities in humans and the environment ( Illumina, 2 015 ). All of the NGS technologies result in millions of reads consisting of short fragments of nucleic acid s (the building blocks of DNA) which need to undergo bioinformatics analysis (computer programs that identify the DNA) in order to determine the a ssociated microbial communities in the sample. Bioinformatics uses a combination of known databases that have to be built , algorithms, computational techniques , and statistical tools to analyze and match the complex genetic and genome sequence data genera ted by NGS technologies to known sequences so that organisms can be identified . The primary aims of bioinformatics include the organization of data for research access and entry, development of tools that help analyze all of the complex data, and the use of these tools to analyze biological data in a meaningful manner (Luscombe et al. , 2001) . Viral metagenomics integrates bioinformatics tools into pipelines in order to analyze and characterize entire viral communities (Aw et al. , 2014 ) . The basic steps 21 du ring bioinformatics analysis of viral metagenomic data include sequence read preprocessing (quality control and trimming), assembly, and annotation (Kunin et al. , 2008) . Common viral bioinformatics tools used for Illumina sequencing reads and their u ses a re described in Table 5. Table 5 . Bioinformatic computational t echn iques and tools used for viral m etagenomics Bioinformatics s tep Description Tool e xamples Quality control check Check raw data to ensure sequence quality FastQC Quality trimming Ad apter removal AdapterRemoval, Cutadapt, and Trimmomatic Assembly Fragmented nucleotide sequences assembled into overlapping segments of nucleic acids (contigs) Velvet and IDBA - UD Annotation Compare contigs with sequence databases to identify genes and assign biological information tBLASTx or BLASTn Taxonomic classification Identification and organization of the virus species present MEGAN Bioinformatic analysis of the viral sequences generated using NGS technology includes an initial quality co ntrol check followed by sequence trimming, assembly, annotation, and taxonomic classification. FastQC is a common control tool used to check the quality of raw sequence data. Quality trimming of the sequences is then performed to trim sequences to desire d lengths and remove contaminant adapters (Lindgreen, 2012) . Next, fragmented nucleotide sequences are assembled into contigs or overlapping segments of nucleic acids in a process called sequence assembly. To date there are numerous metagenomic sequence assembly tools, however these tools can vary in suitability depending on the sequencing technology used. A few 22 examples of the assembly tools currently used for viral metagenomics include Celera software, IDBA - UD, MetaVelvet, and Velvet ( Vázquez - Castellan os et al. 2014, Wylie et al. 2013 ), however Velvet and IDBA - UD are generally used for Illumina sequence analysis (Aw et al., 2014b, Vázquez - Castellanos et al. , 2014, Wylie et al. , 2013) . Once assembled, annotation and taxonomic classification of the assem bled sequence reads is performed. Genome annotation is where elements of the genome are identified (gene prediction) and biological information is linked to specific sequences. For viral metagenomics, the most commonly used method of annotation includes c omparing sequences to Genbank using tBLASTx or BLASTn (Bibby et al. , 2011; Leclerc et al. , 2001; Mokili et al., 2012; Wylie et al., 2013) . Following annotation, viruses can be grouped into taxonomic classifications and phylogenetic trees to analyze the vi rus communities present. To date, the most common software used for analysis of virus communities is the Metagenome Analyzer (MEGAN) (Aw et al., 2014 , Kim et al., 2011, Moore et al., 2015, Park et al., 2011) . 4.2 Metagenomic Insights into the Virome and Fo od Safety Current metagenomic research has provided insights into microbial communities associated with a number of environmental samples. With reference to viral metagenomics, it is suggested that less than 1% of viral diversity has been explored with unknown (novel) sequences ranging between 60 and 99% in human or environmental samples (Mokili et al., 2012) . The first environmental viral metagenomics study investigated viruses in marine waters through shotgun library sequencing (Mokili et al., 2012, R osario & Breitbart, 2011) . The results suggested that most viral community diversity is currently un described and supported the conclusion that the majority of identifiable viruses in marine environments are phages (Breitbart et al., 2002) . Using NGS tec hnology , it has been suggested that not only do marine waters have high viral diversity, 23 but also diversity varies with geographic region and consists of a large proportion of single - stranded DNA viruses (Angly et al., 2006) . Next generation viral sequenc ing of human feces and wastewater revealed a large proportion of phage, many of which belong to ds DNA bacteriophage of the Caudovirales order (Aw et al., 2014, Kim et al. , 2011) . Furthermore, wastewater contained a wide array of human viruses which were dominated by three adenovirus species (B, C and F), E nterovirus B, polyomaviruses and papillomavirus (Aw et al., 2014 ) . In reclaimed water, eukaryotic viral sequences belonging to plant pathogens of agriculturally important crops were dominated by viruses in the ss DNA Geminiviridae and Nanoviridae families , however viruses infecting numerous animal species (vertebrate and invertebrate) were also identified (Rosario et al., 2009) . N ext - generation sequencing technology is now being used as a tool to dete ct and track pathogen out breaks and transmission routes (Bergholz et al. , 2014) . For example, metagenomic sequencing and analysis of fecal samples collected from individuals involved in previous gastroenteritis outbreaks of unknown etiology in New Zealan d were able to identify eight viruses including human enteric adenov irus, rotavirus, and sapovirus (Moore et al., 2015) . In addition NGS is now applied as a diagnostic tool in plant viral disease and in the discovery of insect viruses (Adams et al., 2009, Liu et al. , 2011) . These are just a few examples of how NGS technology has already provided knowledge of the virus world. Recognizing that NGS technology could be used as a tool for monitoring food safety, scientists are now beginning to investigate the bacterial and viral communities associated with foods. To date, a number of studies have focused on the microbial communities associated with fermented foods. In a study by Park et al. ( 2011 ) , the viral ds DNA in fermented k im chi, sauerkraut, and shrimp was amplified and sequenced using Roche 454 pyrosequencing. This 24 study found that 99.3% to 99.9% of viral reads showed the greatest sequence similarity to phages, with 99.9% of phages belonging to bacteriophage s in the C au dovirales order (Park et al., 20 11) . Although fermented food viral communities were less diverse than other environmental habitats, these samples contained a large proportion of unidentified viral sequences suggesting a lack of data and understanding of viral genomes associated with the se samples (Park et al., 2011) . These data advance our current knowledge on the diversity of viruses and ultimately the ecological roles that these viruses play in food systems. To date, metagenomic approaches to study the genet ic material in fresh prod uce and sources of fresh produce contamination have focused primarily on bacterial communities . In a study by Ottesen et al. ( 2013 ) , NGS was used to characterize the tomato microbiome by sampling different parts of the tomato plant (fruit, flowers, leaves , stems, and roots) to identify ecological contributors to Salmonella persistence. This study observed 10 phyla from bacterial, eukaryotic, and viral domains and identified Pseudomonas and Xanthomonas as the most common bacterial taxa across all plant regi ons . Although Salmonella was not detected, this is one of the first studies to investigate the microbiome in fresh produce, concluding that microbial diversity decreases as distance from the soil increases and bacterial diversity v aries between different parts of the tomato plant (Ottesen et al., 2013) . In addition, metagenomics has been used to study the impact of suggested sources of pathogen contamination, including irrigation water, on the microbial surface communities of fresh produce. Telias et al. ( 2011 ) discovered major differences in the bacterial composition between ground and surface irrigation water, however the surface microbial communities of tomatoes irrigated with these waters were not significantly impacted. In a more extensive study, Le ff & Fierer ( 2013 ) u sed NGS technology to study bacterial communities on numerous fruits (grapes, strawberries, apples, peaches, tomatoes) 25 and vegetables (lettuce, spinach, mushrooms, sprouts, peppers) at the point of sale, investigating how community stru cture differs between produce types and if farming practices contribute to composition. This study revealed that bacterial communities: (1) a re highly diverse and vary with produce type , (2) o n average, are similar between produce types grown in similar e nvironments (tree vs. ground) , and (3) differ significantly in composition based on farming practices (conventional vs . o rganic) (Leff & Fierer, 2013) . Metagenomic technology has also been applied to study the change of bacterial communities in spinach d uring packaging and storage. Results suggest that bacterial diversity, richness, and evenness significantly decreased when spinach was package d and stored at 4°C and 10°C with the entire microbiome reduced from 11 to 5 phyla after 1 day of storage at 4°C (Lopez - Velasco et al., 2011) . These studies serve as examples of how metagenomics can enhance our knowledge of microbial communities associated with foods, especially fresh produce, from farm - to - fork and lead to the identification of possible control poin ts for enhanced safety . 5 . Post - Harvest L eafy Green Processing and Viral Contamination 5 .1 Leafy Green Commercial Processing P ractices As stated previously, fresh produce production practices are becoming more intensive to accommodate growing consumer dem and. Following harvest, fresh produce is subject to a number of processing techniques which may vary between location and food type. In addition, many fruits and vegetables are often combined (before and after sale) to make up complex foods or beverages such as salads mixes, smoothies, and sandwiches. The complexity of the food supply chain provides multiple opportunities for human pathogen contamination, making it difficult to determine the single vehicle of transmission or source of contamination. Lea fy green 26 vegetables, including lettuce (iceberg, romaine, red leaf, butter, etc.), escarole, endive, spring mix and spinach, are an example of a food commodity that is commonly consumed raw. Leafy greens are provided to the consumer either as bulk produc ts to be washed (e.g., head of lettuce) or as ready - to - - intended for bulk sale, the entire head is usually cut manually with a harvesting knife (with wrapper leaflets removed), placed onto a pr ocessing platform for packaging, and transported for vacuum (iceberg) or spray - vacuum (romaine) cooling and cold storage. Lettuce may be vacuum cooling followe and shelf - life of porous leafy greens. Following cold storage, lettuce is packed into shipping containers and transported to distribution centers that further transport the food prod uct to retailers or foodservice establishments (CFSAN, 2009 ) . The supply chain for packaged fresh - cut lettuce is even more complex. Practices such as coring, rinsing, and outer leaflet removal at the time of harvest are now implemented in the field in an attempt to provide cleaner, safer lettuce to processing facilities (NFPA, 2001). Following harvest and cooling, lettuce intended for RTE salads undergoes a number of additional processing steps to reduce foodborne pathogen transmission and improve over all food quality. Currently, standard steps in post - harvest leafy greens processing for RTE salads include lettuce shredding, washing, shaker table dewatering, centrifugal drying, and packaging prior to cold storage and transportation to distribution cent ers and end - users. Although due diligence is necessary at all production levels to ensure food product safety, wash water disinfection and monitoring of san i tizer levels during processing are seen as essential to minimize foodborne disease (CFSAN, 2014 ) . To date, numerous physical (ultrasound high 27 pressure, ultraviolet, ionizing radiation) and chemical (chlorine dioxide, sodium chlorite, quaternary ammonium compounds, peroxyacetic acid) disinfection methods have been investigated for use in foods, howev er, washing with a chlorine - based sanitizer is the disinfection method most commonly used in the fresh - cut produce industry due to the low cost (CFSAN , 2014) . Fresh produce is generally placed into a large tank and washed with recirculated water containin g a sanitizer, a process known as fluming. Guidelines for washing fresh produce in chlorinated water include a maximum free chlorine (hypochlorite) concentration of 200 ppm and a 1 to 2 minute contact time (CFSAN , 2014) . Currently, the National Advisory Committee on Microbiological Criteria for Foods uses a 5 - log pathogen reduction performance standard for fruit and vegetable juice production (CFSAN , 2001) . However, this standard is primarily targeted towards bacteria l pathogens since there is no defined criterion for antiviral disinfectants , a 3 - log reduction has been generally accepted for virus efficacy testing (Allwood et al, 2004; Gulati et al. , 2001 ) . 5.2 Post - Harvest Fresh Produce C ontamination and Viral Survival on F oods Human pathogen contamina tion of leafy greens can occur at any stage of post - harvest production including processing, packing, storage, and transportation. Critical control points include the quality of water used (cooling, washing), worker hygiene, and the condition or overall c leanliness of processing equipment, cooling facilities, storage and packaging containers, and transportation vehicles (CFSAN , 2006) . To date, numerous studies have investigated bacterial pathogen survival (primarily E. coli O157:H7) and contamination durin g farm - to - fork production (Beuchat, 2002) . Interestingly, research has shown that bacterial pathogens are easily transferred between lettuce and processing equipment (such as coring knifes and processing lines) and preferentially attach to fresh cut lettu ce surfaces allowing for survival and persistence 28 during processing (Buchholz et al. , 2012a, Buchholz et al. 2012b, Seo & Frank, 1999, Taormina et al., 2009) . To date numerous studies have investigated virus attachment and survival in fresh produce, al l of which affect virus recovery and disinfection. Viruses are thought to use physicochemical forces, specifically electrostatic forces, for nonspecific attachment to solid surfaces such as fresh produce which can be disrupted by a high pH (Deboosere et al., 2012, Vega et al., 2008) . For example Vega et al. (2005) observed maximum adsorption of MS2 to butterhead lettuce at a pH of 3.0, while a basic pH of 8.0 led to almost complete dissociation of the virus from the lettuce surface. This information is critical for understanding how to better recover, remove, and inactivate viruses from lettuce surfaces during processing. Temperature is a well - known factor influencing both the survival and internalization (virus entry into plant interior) of pathogens in lettuce. Regardless of the specific fruit or vegetable, survival studies consistently show that non - enveloped viruses (rotavirus, MS2 phage, poliovirus, adenovirus) are able to survive for long periods of time (25 - 76 days) at traditional storage temp eratures (4°C) (Badawy et al. , 1985 , Dawson et al., 2005, Ward & Irving, 1987) . These data stress the importance of limiting viral pathogen contamination of fresh produce at the field level prior to harvest and storage. Changes in storage conditions such as increased temperatures and CO 2 levels have shown to significantly decrease virus survival in fresh produce and the time of survival has also been shown to vary with the food commodity and virus type (Dawson et al., 2005, Rzezutka & Cook, 2004) . For ex ample, Allwood et al. (2004) found that the decimal reduction time (the time needed at a given temperature to kill 90% of the organisms) for MS2 on iceberg lettuce was reduced from 5 days at 4°C to 3 days at 37°C. Interestingly, studies addressing multipl e food commodities have observed greater virus survival in lettuce 29 compared to other fresh produce commodities (Badawy et al., 1985, Croci et al. , 2002) . Currently there is limited knowledge as to why virus survival is greater on lettuce in comparison to other food commodities, however suggestions include protection from the rough surface of lettuce, resistance to naturally occurring antimicrobials, or protection due to internalization (virus entry into plant interior) through roots or cut surfaces (Badawy et al., 1985, Seymour & Appleto n, 2001, Wei et al. , 2010) . Despite survival study insights into virus persistence in the food supply chain, there is currently limited research investigating where in the supply chain contamination occurs. There have be en numerous studies investigating bacterial pathogen reduction on fresh produce using chlorine sanitizers. Pilot - scale studies have found that chlorine - based sanitizers generally reduce bacterial pathogen populations on lettuce between 1 and 3 logs (David son et al., 2013, Gil et al. , 2009) . Laboratory studies investigating the effects of chlorine on viruses when inoculated onto fresh produce have shown that viral (MS2, feline calicivirus , murine norovirus ) reduction generally does not exceed 3 logs when e xposed to a variety of free chlorine (15 - 800 ppm) levels (Allwood et al., 2004, Dawson et al., 2005, Fraisse et al., 2011, Gulati et al., 2001 ) . In a study by Fraisse et al. (2011), feline calicivirus, murine norovirus, and hepatitis A virus populations o n inoculated lettuce decreased 1.9, 1.4, and 1.4 logs, respectively, when washed for 2 min in 15 ppm of free chlorine . This can be compared to an overall 1 log reduction using water alone and a 3.2 (feline calicivirus), 2.4 (murine norovirus), and 0.7 (he patitis A) log reduction using 100 ppm of a peroxide - based disinfectant (Fraisse et al., 2011) . In another study, than strawberries (1 - 1.2 log) when exposed to 20 ppm free chlorine for 3 to 5 min (Casteel et al. , 2008) . Currently, there are a limited stu dies investigating bacterial and viral reduction on fresh 30 produce during simulated commercial processing. Davidson et al. (2013) showed that E. coli O157:H7 populations on lettuce were not significantly reduced when washed with either water or 30 ppm of fr ee chlorine during simulated commercial processing. In addition, the use of 30 ppm of available chlorine in the wash water at both a pH of 7.85 and 6.5 were found to result in significant population reductions on the lettuce (Davidson et al., 2013) . In a v iral study by Casteel et al. 2009, an industrial - scale processing unit consisting of a washing compartment, grates, and conveyor with tap water spray was used to study chlorine inactivation of MS2 on strawberries. This study found that processing strawberr ies with wash water containing 20 and 200 ppm free chlorine inactivated 92% and 96% (~1 log PFU) of the MS2, respectively, compared to 68% using water alone. More research is needed to provide a better understanding of virus inactivation on fresh produce during simulated commercial processing. 31 II. RESEARCH GOALS AND OBJECTIVES This thesis wa s split into two primary studies and ha d a relat ed set of goals and objectives. The first study goal ( Part III ) was to use NGS technology and metagenomi c techniques for the first time to identify the virus communities (virome) present in irrigation water and lettuce and use this information to better understand contamination in the field . The spec ific objectives were to : 1) Evaluate a method of virus recov ery from lettuce 2) Identify and evaluate the diversity of virus communities present in irrigation water and lettuce (iceberg and romaine ) at the field level The goal of the second study (Part IV) was to i nvestigate the efficacy of current post - harvest le afy green processing and disinfection practices to better understand viral risks from farm - to - fork during a contamination event . Specifically, the goal was to a ssess the efficacy of a chlorine - based sanitizer against coliphage MS2 ( an enteric virus surrog ate ) on romaine lettuce during simulated commercial processing . The s pec ific objectives were to: 1) Evaluate MS2 reduction on romaine lettuce during and following small - scale commercial leafy green processing (shredding, flume washing, shaker table dewateri ng , centrifuge drying ) with and without a sanitizer wash treatment 2) Determine MS2 levels in the flume wash water and centrifug ation water following processing 32 II I . METAGENOMIC IDENTIFI CATION OF VIRUS COMM UNITIES ASSOCIATED WITH LETTUCE AND IRR IGATION WAT ER 1. Introduction New scientific methods and genomics tools can help us take a broad view of food safety like never before, particularly for hazards such as viruses where traditional methods have limited our ability to monitor. F ood safety monitoring ef forts in the United States food supply chain (Crutchfield & Roberts, 2000) . However d espite recent advances in monitoring , f oodborne illness remain s a serious threat to public health with approximately 48 mill ion people in the United States affected each year (CDC, 2014). In addition fresh produce, which is commonly consumed raw, is becoming increasingly recognized as a vehicle of human pathogen transmission. Specifically, leafy green vegetables were the food commodity responsible for the highest number of foodborne illnesses (2.2 million) between 1998 and 2008 in the United States (Painter et al., 2013). Recent foodborne outbreaks also suggest that viruses play a larger role than previously thought and stu dies on norovirus in particular have shown strong associations with leafy green consumption (Gould et al. 2013) . Viruses are host - specific, obligate intracellular infectious agents that are ubiquitous in nature and can cause a wide range of diseases. Exa mples of viral hosts include bacteria (bacteriophage), algae, plants, animals (invertebrate and vertebrate), and humans. Human enteric viruses, which are commonly transmitted through the fecal - oral route, are of particular public health concern. These vir gastrointestinal tract and are shed at extremely high concentrations into the environment through the feces or vomit of an infected individual . Food and water can then become contaminated at both pre - and post - harvest levels of production. Common examples of food and waterborne 33 viruses include adenoviruses , astrovirus es, enteroviruses, hepatitis A and E virus es , norovirus, reoviruses and rotaviruses which can cause gastroenteritis, meningitis, liver disease, in fantile diarrhea, respiratory illness, or neurological symptoms ( Cook, 2013, Fong and Lipp, 2005, Koopmans et al. 2004 ). The microbial quality of irrigation water is now recognized as one of the primary pre - harvest factors influencing the microbial quali ty of fresh produce. Currently, there is no universal standard for irrigation water microbial quality and traditional bacterial indicators of fecal pollution (coliforms, E. coli , enterococci) fail to identify enteric virus hazards in water ( Harwood et al. , 2005 ) . In addition, surveillance and on - sight field investigations of the environmental sources of fresh produce contamination, such as irrigation water, during foodborne viral outbreaks is limited by current methods for virus detection. To date the re lationship between irrigation water and fresh produce viral contamination is poorly understood and greater assessments are needed to provide resolution on the types of viruses present. Current methods for virus detection include cell culture, electron m icroscopy, PCR, and metagenomics. Although cell culture is still one of the most often used methods today, virus detection is difficult due to the host specificity of viruses which requires the correct cell line for proliferation and isolation. Another m ajor disadvantage is that many viruses such as human noroviruses are currently unable to grow in any of the known cell lines. Molecular methods such as PCR , RT - PCR (for the detection of RNA viruses), and qPCR are sequence - dependent detection methods whic h have allowed for virus detection and quantification but first require knowledge of the viral nucleic acid (DNA or RNA ) sequence and therefore discovery of novel viruses is not possible (Mokili et al. 2012). An exciting emerging field of science and tech nology includes metagenomics and next generation sequencing. 34 Metagenomics is a sequence - and culture - independent approach for studying entire microbial communiti es in an environmental sample using NGS technology. Current NGS tools include Roche 454 G en ome S equencers, Illumina (Genome A nalyzer, Hiseq, Miseq), and the Applied Biosystems (AB) SOLiD sequencing platforms which use pyrosequencing, sequencing - by - synthesis, and sequencing - by - ligation technology, respectively. These parallel sequencing technol ogies result in millions of reads per run and have led to significant cost reductions. To date, viral metagenomic research using Roche 454 or Illumina sequencing technology has primarily focused on the virus communities in water (marine, reclaimed, and wa stewater) and human feces ( Angl y et al., 2006, Aw et al., 2014, Kim et al. 2011, Mokili et al. 2012, Rosario et al., 2009 ) . Data suggests that less than 1% of vi ral diversity has been explored and a large proportion of viruses in both water and clinical s amples belong to double - stranded DNA bacteriophage of the Caudovirales order. In addition, NGS is now being used as a tool to detect and track viral pathogen out breaks and transmission routes (Moore et al. 2015) . These are just a few examples of how NGS t echnology has already provided knowledge of the virus world. Recognizing that NGS technology could be used as a tool for monitoring food safety, scientists are now beginning to investigate the viral communities associated with foods such as fermented k im ch i, sauerkraut, and shrimp ( Park et al. 2011 ). Such studies can help advance our current knowledge on the diversity of viruses and ultimately the ecological roles these viruses play in our food systems. To date, current application of metagenomic approach es to study the genet ic material in fresh produce and sources of fresh produce contamination have focused primarily on the bacterial communities ( Leff & Fierer 2013, Lopez - Velasco et al. 2011, Ottesen et al. 2013, Telias et al. 2011 ). There is currently l imited research on the virus communities associated with fresh produce and how the food production chain affects the viral microbiome. 35 Applying N GS technology to studying the virus communities associated with irrigation water and fresh produce is a promis ing tool for enhancing knowledge of viruses in our food systems and can lead to the identification of possible control points for enhanced safety . 2. Research G oals and Objectives The goal of this portion of the thesis wa s to use NGS technology and metage nomic techniques for the first time to identify the virus communities (virome) present in irrigation water and lettuce and use this information to better understand contamination in the field . The specific objectives were to: 1) Evaluate a method of virus recovery from lettuce 2) Identify and evaluate the diversity of virus communities present in irrigation water and lettuce (iceberg and romaine) at the field level 3. Materials and Methods 3.1 Virus Recovery from Lettuce 3.1.1 Bacterial Host and Bacteriophage Preparation Bacter iophage and respective bacteria host s were prepared as seed cultures following a standard procedure for plaque assay s ad apted from EPA Method 1602 (EPA, 2001). The bacteriophage used in this study included P22 ( provided by Dr. Charles Gerba, University of Arizona , AZ, USA ) and F+ specific coliphage MS2 (ATCC#15597 - B1 , ATCC, Manassas, VA, USA ) which infect Salmonella (LT2 pLM2 1217 HER #1023 , Félix d'Hérelle Reference Center for bacterial viruses of the Université Laval , Quebec, Canada ) and Escherichia coli ( E. coli Famp ATCC#700891 , ATCC ) hosts, respectively . Salmonella LT2 and E. coli Famp host stock cultures were prepared by rehydrating lyophilized cultures in tryptic soy broth (TSB , Becton, 36 Dickinson and Company, Sparks, MD, USA ) and incubating overnight at 37°C before adding 10 to 20% glycerol by volume. A 1 - mL aliquot was then added to a cryovial and stored at - 80°C. Escherichia coli Famp was prepared in TSB containing a 1% v olume /v olume (v/v) solution of ampicillin - streptomycin pr epared by dissolving 0.15 g ampicillin sodium salt ( Sigma - Aldrich, St. Louis, MO, USA ) and 0.15 g streptomycin sulfate ( Sigma - Aldrich ) in 100 mL reagent grade water, which was then filtered through a 0.22 µm filter and stored in 5 - mL vials at - 20°C. A w orking stock solution of b acteriophage P22 or MS2 was prepared by first rehydrating lyophilized phage in TSB and then diluting in phosphate buffered saline (PBS). A double - agar overlay method was used for phage replication and enumeration. To perform the d ouble a gar o verlay, 1 mL of bacterial host grown to log phase (see below) and 1 mL of each phage suspension (dilutions) were added to 2.5 mL, 1.5% trypticase soy agar (TSA , Becton, Dickinson and Company ) overlay s ( tempered in a 48 o C water bath). After addition of host and sample the tube was gently mixed by rolling the tube between the hands and immediately pour ed onto solidified TSA plates for each dilution series. Plates were allowed to solidify, inverted to avoid condensation and then incubated at 3 7°C for 24 h to allow the growth of the monolayer of bacterial host cells and plaques formation ( areas where the phage replicated and lysed the bacteria l cells ) . Plates with high plaque counts (exhibiting a lacey pat tern with approximately 1000 Plaque Fo rming Units (PFU) /plate) were flooded with ~10 mL of TSB and incubated with shaking at 4°C for 1 hr. Finally, the TSB was recovered using a pipette and bacteria were removed by filtering through a 0.22 µm filter . The resulting P22 (10 10 PFU/mL) and MS2 ( 10 9 - 10 10 ) bacteriophage working stock cultures were stored at 4°C . Bacteriophage working stocks were then used to maintain future working stocks. To maintain P22 and MS2 working stocks, overnight log phase culture s of Salmonella LT2 and 37 E.coli Famp we re prepared by adding 1 mL of host stock culture from the freezer to 9 mL of TSB followed by overnight incubation at 37°C. Log - phase host cells were prepared by inoculating 1 - mL of the overnight host into a desired working stock volume (500 mL) followed by 4 h of incubation at 37°C with shaking . Once the bacterial host cells reached log phase, 1 mL of bacteriophage working stock was added to the TSB bacterial suspension and the mixture was incubated overnight at 37° C . E . coli Famp was grown in TSB conta ining a 1% v/v soluti on of amplicillin - streptomycin, as described for host stock preparation . The majority of the bacteria were infected and lysed by the phage, releasing virus into the broth. Afterwards , the sa mples were filtered through 0.45 µm and the n through 0.22 µm filters to remove bacteria and the new working bacteriophage stocks were stored at 4°C (10 9 - 10 10 PFU/mL ) . T o prepare bacterial hosts for the plaque assay, 1 mL of host stock culture from the freezer was added to 9 mL of TSB and incubat ed overnight at 37°C. Log - phase host cells were prepared by inoculating 1 - mL of the overnight host into 30 - 40 mL of TSB and incubating for 4 hr at 37°C. Again, E . coli Famp was grown in TSB containing a 1% v/v solution of amplicillin - streptomycin, as des cribed for host stock preparation . 3.1.2 Lettuce Inoculation, Elution, and Plating of Bacteriophage B acteriophage recovery experiments were performed to determine the effectiveness of virus elution from lettuce . Both bacteriophage MS2 and P22 w ere used a s inoculants for this pilot recovery assay. Unpackaged romaine lettuce heads were purchased from local grocery stores and stor ed at 4°C for a maximum of 48 h . On a ste rile surface, 50 g of the outer leaflets cut with a scalpel 2.5 to 5 cm from the core w as inoculated with 1 mL of P22 or MS2 diluted to 10 4 PFU/mL by hand - pipetting droplets (60 - 70) from a 1000 µL pipette evenly onto the leaflet surfaces in a biosafety cabinet. For the inoculation suspension, bacteriophage stock containing 38 approximately 10 9 - 10 10 PFU/mL was diluted 10 - fold to 10 4 PFU/mL in 1x Phosphate Buffered Water (PBW). Inoculated leaflets were then allowed to air - dr y at room temperature for 20 min (following procedure by Dubois et al., 2007 ) . A ~50 - g, un inoculated lettuce sample was pr ocessed as a negative control . Inoculated lettuce samples were then eluted. E ach sample was placed into a Whirl - pak bag (Nasco filter Whirl - pak 19 x 30 cm , Fort Atkinson, WI, USA ) and soaked in 250 mL of 100 mM Tris (UltraPure Tris, Invitrogen, Carlsbad , CA, USA) 50 mM glycine (Tris - glycine ) elution buffer at pH 9.5 , a method adapted from Dubois et al., 2007 . A hi gh pH was used to disrupt the electrostatic forces that viruses use for surface attachment (Vega, Garland, & Pillai, 2008) . The samples were then placed on a rocking platform shaker ( Model 100 , VWR, Radnor, PA, USA) a t full speed for 20 min . Samples were inverted half - way through the shaking process (10 min) to soak both sides of the lettuce leaflets equally . The solution was then recovered and transferred to a sterile 500 mL plastic container (Nalgene wide mouth environmental sample bottle with lid) while recording the recovery volume. Then 6 .0 M HCl was use d to adjust the pH of the eluent to 7.2 ± 0.2 using a pH probe disinfected between s amples by immersion in 10% bleach for 10 to 15 min and neutralized with sterile 5% sodium thiosulfate . The eluent solution was then vortexed, diluted 10 - fold in PBW, and plated both undiluted and at a 10 - 1 dilution (2 mL sample and 0.5 mL host) on TSA pla tes using a double agar overlay method adapted from EPA Method 1602 (EPA, 2001) as described above. Positive (spot plate), negative (2.5 mL of host), TSA media, overlay, and PBW controls were included in each experiment. In addition, the 10 - 8 , 10 - 9 , and 1 0 - 10 dilutions from the inoculation suspension were assayed by the same d ouble a gar o verlay method to confirm the initial concentration upon inoculation and determine percent 39 recovery. A fter the overlay, plates were al lowed to solidify for 10 min , inverted and incubated at 36 ± 1.0°C for 16 - 24 h prior to enumeration of plaques in the monolayer of the bacterial lawn. 3.1.3 Plaque Concentration Assay and Percent Recovery Calculation To calculate the p hage concentration, the plaques were counted on TSA overla y plates within each dilution series containing a countable range (10 - 300 PFU). After choosing the appropriate dilution, the number of plaques were summed for each of the three plates and divided by the volume of sample on each plate (2 mL ). This value wa s divided by the number of plates with plaques in the countable range and then multiplied by the dilution factor to obtain average PFU/mL in the eluent as shown in Formula 1 . If two dilutions were in the coun table range for an individual sample, the PFU/m L was calculated within each dilution series then averaged to obtain a single plaque concentration (value used in percent recovery calculation) . To calculate the concentration of phage per g of lettuce (PFU/g), the calculated p hage concentration from Formu la 1 was multiplied by the eluent volume recovered and divided by the lettuce weight of each sample, as shown in Formula 2. The estimated initial phage concentration was calculated by multiplying the inoculum concentrat ion (PF U/mL) by the volume of inocul um concentrate added (1 mL) and dividing by the total weight of lettuce added to the suspension, as shown in Formula 3. The formula used to calculate the percent recovery of phage from the virus elution recovery protocol is shown in Formula 4. (1) 40 3.2 Virus C ommuniti es from Lettuce and Irrigation W ater 3.2.1 Irrigation Water and Lettuce Sample C ollection Lettuce and irrigation water samples were collected in Yuma, AZ, during December of 2013 . Irrigation water was sampled from six sites along the Yuma main canal . At e ach location temperature, pH, conductivity and turbidity were measured using two portable meter s (HACH HQ40 d Portable pH, Conductivity , DO Multi Parameter Meter, and HACH 2100Q Portable Turbidity Meter respectively , Loveland, CO, USA ). Images of the Yuma irrigation water sampling locations are p rovided in Figure A1, Appendix . Sample collection consisted of lowering a 5 - L disinfected bucket into the irrigation canal from an overpass or other accessible site , avoiding contact with the shoreline or bottom se diment to minimize turbidity . Prior to use , the bucket was sanitized by exposing to 10% bleach for 10 to 15 min and neutralizing with 5% sodium thi osulfate for a maximum of 5 min . For each irrigation water sample , five 20 - L collapsible containers (Cole - Pa rmer , Vernon Hills, IL, USA ) were filled for a total of 100 L . In addition, 500 mL of water were collected for Colilert and Enterolert testing to detect the presence of the indicator bacteria E. coli and enterococci , respectively. Colilert and Enterolert testing was performed using a Quanti - Tray/2000 test kit (IDEXX laboratories Inc . , Westbrook, ME, USA). C 41 Briefly, the water sample was vortexed and 100 mL was transferred to a sterile graduated bottle. The reagent provided in the kit was then added to th e sample and mixed before the entire volume was poured into a Quanti - Tray/2000. Once sealed, the samples were incubated at 41 ± 0.5°C and 35 ± 2.0°C for the Enterolert and Colilert tests , respectively. Yellow wells ( total coliform) and ultraviolet fluor escent wells (enterococci and E. coli ) were counted and the IDEXX most probable number (MPN) generator (version 3.2) was used to obtain the MPN/100mL . Enterolert testing included Nanopure water, Streptococcus bovis and E.coli as negative controls and Ente rococci faecium as the positive control (IDEXX - QC Enterococci). Colilert testing included Nanopure water, Pse u d omonas aeruginosa , and Klebsiella pneumonia as negative controls and E. coli as the positive control (IDEXX - QC Coliform and E. coli ). A total o f 42 (21 i ceberg and 21 r omaine ) lettuce heads were collected at different stages of farm - level production. A total of 8 iceberg and 8 romaine samples were hand cut at ground level by the research crew using gloves and a sterile harv esting knife as a cont rol. The outer leaflets were removed before placing the heads in Whirl - pak bags. A total of 8 iceberg and 5 romaine lettuce samples were hand cut by workers and placed on a packaging machine for bagging prior to sampling . In addition, 5 iceberg lettuce s amples that were harvested and bagged by workers were collected following a 30 min worker break. Finally, a total of 8 samples were collected post worker chop and wash (n=3) and from buckets containing mixed romaine head lettuce samples to be packed as sa lad (n=5). Chop and wash is a harvesting process used in bagged salad production in which the core of the romaine head is removed and the lettuce leaflets are conveyed and rinsed on a processing machine in the field prior to collection in large storage co ntainers . Lettuce leaflets will undergo further post - harvest processing (shredding and disinfection) prior to packaging and sale. 42 3.2.2 Virus Concentration and Purification Viruses in irrigation water samples were concentrated using a tangential flow, l ow cost, disposable hollow fiber ultrafiltration system (Asa hiKasei, Toky o, Japan) (Figure A3, Appendix ) . To prime the system, 1 L of 0.01% sodium polyphosphate (NaPP) solution was recirculated through the ultrafilte r at a rate of 17 00 mL /min for 15 min. Samples were then added to the reservoir and pumped at 2900 mL /min. The n the flow regulator was adjusted to a flow rate of 1200 - 1300 mL /min with the system pressure maintained below 10 psi until the final concentrate was between 250 and 300 mL . In additi on, 300 mL of surfactant solution (0.01% NaPP, 0.5% Tw een 80, and 0.001% Antifoam) was circulated at a rate of 600 mL /min for 5 min and combined with the sample for a final volume of about 500 mL . The final concent rate was then shipped on ice to the labora tory at MSU for further processing. The same elution procedure was used as described in P art III, S ection 3.1.2, page s 37 - 3 8 . Each ~50 - g lettuce sample was eluted with 250 mL of buffer at pH 9.5. Following pH adjustment using 1 .0 M HCl to 7.2 ± 0.2 , 3 .0 M NaCl was added to the recovered solution (volume varied with recovery) for a final concentration of 0 .3 M NaCl. The same procedure was used to adjust the final concentration of irrigation water samples to 0.3M NaCl . After thorough mixing , the sample w as cooled for 30 min at 4 °C. Polyethylene glycol precipitation was then used to further concentrate viruses in irrigation water and lettuce samples . Molecular biology grade PEG 8000 powder (Promega Corporation, Madison, WI , USA) was gradually added to ea ch sample and mixed thoroughly for a final concentration of 10% (w/v). After 18 h of incubation at 4°C the sample was centrifuged at 10,8 00 x g (8000 rpm) for 30 min at 4° C ( Beckman Coulter JS - HS centrifuge , Brea, CA, USA ) . The resulting pellet was dissolv ed in 20 mL of PBS at room temperature for 1 hr. The walls of the container were rinsed before adding 43 and mixing an equal volume of chloroform (10 - 20 mL) to the PBS to remove the PEG and purify the sample. The solution was then vortexed gently for 30 s a nd centrifuged at 3000 x g (4300 rpm) for 15 min at 4° C to collect the supernatant containing virus particles. The remaining supernatant was then passed through 0.45 and 0.22 filters and the final concentrates were stored at - 80°C until further concen trated by Amicon centrifugal ultrafiltration. 3.2.3 Final C oncentration , P urification , and Nucleic Acid E xtraction Following filtration, the concentrate was added to an Amicon Ultra 30 kDA centrifu gation co lumn (Millipore, Billerica, MA , USA ) a nd centrif uged at 1000 x g at 4° C in a swinging - bucket rotor until about 1 mL of sample was left in the filter (~5 - 7 min). The sample was removed from the reservoir and the filter was rinsed with 1.5 mL of the filtrate, vortexed, and added to the final concentrate. Samples were treated with DNase - I (Roche , Indianapolis, IN, USA ) before nucleic acid extraction to remove free nucleic acids from the concentrated virus samples. Viral nucleic acid s w ere extracted using a QIAGEN QIAamp MinElute Virus Spin Kit , following the instructions (QIAGEN, Maryland, USA) . Following extraction, the samples were checked for 16s contamination by PCR using a 50 µL cocktail consisting of 25 µL GoGreen master mix (Promega) , 1 µL - AGAGTTTGATCCTGGCTCAG - - GGTTACCTTGTTACGACTT - 18 µL of water, and 5µL of template. PCR conditions include d a 95° C denaturation for 5 min, followed by 30 cycles at 94 °C for 30 sec, 55°C for 30 s, 72°C for 1 min, and a final extension at 72°C for 10 min. 3.2.4 Random A mplification Viral nucleic acid was amplified using a protocol adapted from Wang et al., ( 2003 ) . In this method, t wo rounds of enzymatic reactions were used to rando ml y amplify the viral nucleic acid. In the first round, two cycles of first strand cDNA synthesis were performed with reverse 44 transcriptase SuperScript III (Invitrogen) and 40 pmol/µL primer A (GTT TCC CAG TCA CGA TCN NNN NNN NN , Eurofins Genomics, Huntsville, AL, USA ). For the first cycl e , 5 µL of template was added to 1 µL of primer A and 4 µL of RNase - free H 2 0 (QIAGEN) for a 10 µL reaction and incubated at 65°C for 5 min followed by 5 min at room temperature. The second cycle of first strand synthesis consisted of a 2 0 µL reaction conta ining 4 µL RT buffer (5x , Invitrogen ), 0.5 µL of RNA - free H 2 0, 1.5 µL of DTT (0.1 M , Invitrogen ), 1 µL dNTP (10 mM , Promega ), 1 µL RNAse OUT (Invitrogen) , 2 µL of SSIII Reverse Transcriptase (Invitrogen) , and 10 µL of the first cycle template. PCR conditio ns for the second cycle included 50 °C incubation for 60 min, followed by 94°C for 2 min and a 10°C hold for 5 min. In additi on, second strand synthesis was performed using Sequenase (bacteriophage T7 DNA polymerase) (A ffymetrix, Santa Clara, CA, USA). Th e Sequenase enzyme can be used for dideoxy - sequenc ing and is useful in that it is not im peded by secondary structures and a llows strand displacement and lacks exonuclease activity. Ten microliters of Sequenase mix consisting of 2 µL of 5x Sequenase buffer , 7.7 µL of RNA - free H 2 0, and 0.3 µL of Sequenase were added for a total reaction volume of 30 µL . PCR conditions for second strand synthesis include d a n 8 min ramp (temperature cycling over a 8 min time period) from 10°C to 37°C , a 37°C hold for 8 min, rapid r amp to 94°C for 2 min, and a 10°C hold for 5 min during which 1.2 µL of diluted Sequenase (1:4) was added. Th e samples were then r amp ed to 37°C for 8 min, held at the same temperature for an additional 8 min, incubated at 94°C for 8 min , and cooled to 10°C to complete the first round PCR template preparation. The second round used the previously generated template and Primer B (GTT TCC CAG TCA CGA TC , Eurofins Genomics ) to amplify the viral nucleic acid. Six microliters of first round template was added to 10 µL of 10X PCR buffer, 2 µL of dNTP (10 mM), 1 µL of 100 pmol/µL Primer B, 6 µL of MgCl 2 , 80 µL H 2 0, and 1 µL of Amplitaq Gold for a 100 µL 45 reaction volume (Applied Biosystems, Austin, TX, USA) . Random amplification PCR conditions included a 95 °C incubation for 15 min, followed by 40 cycles of 94°C for 30 s, 40°C for 30 s, 50°C for 30 s , and 72° C for 1 min. For each sample, three s econd round PCR reactions were amplified and combined for purification. Each PCR run was confirmed by running 5 µL of PCR product on a 1 - 2% agarose gel in which a visible smear of DNA appeared between 500 base pair s ( bp ) to 1 k ilo b ase (kb). A negative control was included for the entire amplification process. The amplified products were purified using a Promega Wiza rd SV Gel and PCR Clean - 3.2.5 S equencing and B ioinformatics Sequencing was performed at the Research Technology Support Facility at Michigan State University. For irrigation water sampl es, libraries were prepared using an Illumina TruSeq kit (Illumina, San Diego, CA, USA) while lettuce sample libraries were prepared using a Rubicon Tenomics ThruPLEX DNA - seq kit (Rubicon Tenomics, Ann Arbor, MI, USA) . Both sample types were sequenced usi ng Illumina HiSeq 2500 Rapid Run flow cell in a 2x100 bp paired end format. Following sequencing, FastQC was used to check the quality of the Illumina sequencing read s (Babraham Bioinformatics, 2015 ). Cutadapt software was used for Primer B removal ( Cuta dapt, 2015 ). Cutadapt parameters included a maximum error rate of 0.2 and minimum overlap of 10 bases. Trimmomatic was used for sequencing adapter removal and quality trimming with the following parameters included: a maximum mismatch count value of 2 all owed for a full match (seed mismatch), a palindrome clip threshold of 30, a simple clip threshold of 10, a minimum adapter length of 8 with both the forward and reverse read kept, removal of low quality leading and trailing bases below a quality of 3, a 4 - base sliding window scan that cuts when the average quality is below 15, and removal of reads less than 30 bases long 46 (Bolger, Lohse, & Usadel, 2014) . Khmer script, interleave - reads.py was used to interleave the paired - end reads prior to assembly ( Crusoe e t al. 2014 ). Assembly was performed using Iterative D e Bruijn Graph A ssembler (IDBA - UD , Hong Kong, China ) software which is an algorithm based on De Bruijn Graph de novo assembly and is used for sequencing data with short reads and uneven sequencing depth (Peng, Leung, Yiu, & Chin, 2012) . Khmer script, extract - long - sequences.py was used to extract contigs larger than 200 base pairs ( Crusoe et al. 2014 ). The Biopieces analyze assembly (Danish Agency for Science, Technology and Innovation) tool was used to analyze the N50, maximum, minimum, mean, total, and number of conti gs for each sequence assembled (Biopieces, 2015 ). The N50 is a statistical measure and is defined as the contig length where using equal or longer contigs produc es half the bases of the gen ome (Biopieces, 2 0 15 ). Assembly results were analyzed using the protein database Basic Local Alignment Search Tool ( BLASTx ) against the National Center for Biotechnology Information (NCBI) Viral Reference Sequence (RefSeq) database. BLASTx paramaters incl uded an expected value (E - value) of < 10 - 5 . The E - value is a parameter describing the number of hits that are expected by chance when searching a database. The BLASTx hits were assigned to NCBI taxonomy using the MEGAN program (version 5.10.0) with the f ollowing parameters for the Lowest Common Ancestor algorithm: minimum score 50, top percent 10 , and min imum support 1 (Huson et al. , 2007) . 4. Results 4.1 Virus Recovery Efficiency from L ettuce Raw data, the calculated eluent, lettuce, and inoculated ph age concentrations used for calculating percent recovery for P22 a nd MS2 are presented in Tables 6 and 7 , respectively. The calculated phage concentration in the eluent (PFU/mL), eluent volume reco vered, and weight of 47 lettuce were used to calculate the ph age concentration on the lettuce. For trials where lettuce was inoculated with 1 mL of P22 at 10 4 PFU/mL, final lettuce phage concentrations ranged from 4.4 x 10 2 PFU/g to 6.4 x 10 2 PFU/g with an l average concentration of 5.4 x 10 2 PFU/g for 3 trials. MS2 phage concentrations on romaine lettuce ranged 1.5 x 10 2 PFU/g to 6.2 x 10 2 PFU/g with an average of 4.2 x 10 2 PFU/g for the 3 trials. The inoculated phage concentration (PFU/g) was calculated by multiplying the inoculum concentration (PFU /ml) by the volume of inoculum applied and dividing by the lettuce weight. The eluent calculated PFU/g was then divided by the inoculum calculated PFU/g to de termine percent recovery. P hage recovery f rom romaine lettuce ranged 43.3% - 77.4% and 2.8% - 91.1 % for P22 and MS2 respectively. The average percent recovery for a ll P22 and MS2 inoculated lettuce trials was 57.9% and 39.0 % respectively. 48 Table 6. Raw data and calculated eluent concentrations, lettuce concentrations, and percent recovery for P22 from inoculated romaine lettuce* Trial Replicate Lettuce weight (g) Eluent volume (mL) Dilution Plate 1 plaque count Plate 2 plaque count Plate 3 plaque count Eluent plaque concentration (PFU/mL)** Eluent plaque concentration s ample average (PFU/mL) Phage concentration on lettuce (PFU/g)** Inoculated phage concentration on lettuce (PFU/g)** % Recovery ** 1 1 48.15 240 .00 10 0 152 194 186 8.9x10 1 1 .0x10 2 5 .2x10 2 1 .0x10 3 51.5 10 - 1 30 18 24 1 .2x10 2 2 51.05 240 .00 10 0 245 235 260 1 .2x10 2 1.3x10 2 6.2x10 2 9 .5x10 2 64. 5 10 - 1 20 29 34 1.4x10 2 3 52.27 240 .00 10 0 199 178 201 9 .6x10 1 1.0x10 2 4 .7x10 2 9.3x10 2 50.1 10 - 1 24 23 17 1.1x10 2 2 1 53.15 240 .00 10 0 174 241 214 1 .1x10 2 1.3x10 2 5 .9x10 2 1 .0x10 3 57.9 10 - 1 31 27 36 1 .6x10 2 2 53.1 1 240 .00 10 0 219 237 207 1.1x10 2 9.8x10 1 4.4x10 2 1 .0x10 3 43.3 10 - 1 22 15 14 8.5x10 1 3 48.09 240 .00 10 0 217 229 217 1.1x10 2 1.1x10 2 5.3x10 2 1.1x10 3 46.6 10 - 1 23 19 18 1 .0x10 2 3 1 50.41 242 .00 10 0 150 163 156 7.8x10 1 9.8x10 1 4.7x10 2 7.9x10 2 59.7 10 - 1 20 26 25 1 .2x10 2 2 48.43 240 .00 10 0 210 220 232 1 .1x10 2 1 .3x10 2 6 .4x10 2 8.2x10 2 77.4 10 - 1 29 21 38 1 .5x10 2 3 54.52 240 .00 10 0 193 202 188 9.7x10 1 1.2x10 2 5.9x10 2 8 .4x10 2 69.9 10 - 1 28 33 20 1 .4x10 3 Avg NA*** NA*** NA*** NA* ** NA*** NA*** NA*** NA*** 1.1x 10 2 5.4x 10 2 9.5x 10 2 57.9 *Inoculum concentrate was 4.9 x 10 4 PFU/mL , 5.4 x 10 4 PFU/mL , and 4.0 x 10 4 PFU/mL for samples Trials 1, 2, and 3 respectively ** Calculated using equations in Part III , S ection 3.1.3, pages 39 - 40 ** * NA: not applicable 49 * Inoculum concentrate was 2.6 x 10 4 PFU/mL, 7.4 x 10 5 PFU/mL, 3.4x 10 4 PFU/mL for sam ples Trials 1, 2 , and 3 respectively ** Calculated using equations in Par t III , S ection 3.1.3, pages 39 - 40 *** NA: not applicable Table 7 . Raw data and calculated eluent concentration, lettuce concentration, and per cent recovery of MS2 from inoculated r omaine lettuce* Trial Replicate Lettuce weight (g) Eluent volume (mL) Dilution Plate 1 plaque count Plate 2 plaque count Plate 3 plaque count Eluent plaque concentration (PFU/mL)** Eluent plaque concentration sample average (PFU/mL) Phage concentration on lettuce** (PFU/g) Inoculated phage concentration on lettuce** (PFU/g) % Recovery ** 1 1 51.02 237.00 10 0 5 7 70 66 3.2 x10 1 3.2 x10 1 1.5 x10 2 5.1 x10 2 29.1 2 45.00 235.00 10 0 78 67 61 3.4 x10 1 3.4 x10 1 1.8 x10 2 5.8 x10 2 30.8 3 48.36 235.00 10 0 79 61 80 3.7 x10 1 3.7 x10 1 1.8 x10 2 5.4 x10 2 32.9 2 1 50.00 231.00 10 0 171 184 179 8.9 x10 1 1.3 x10 2 5.9 x10 2 1.5 x10 4 4.0 10 - 1 34 33 32 1.7 x10 2 2 50.79 231.00 10 0 177 180 191 9.1 x10 1 1.3 x10 2 5.9 x10 2 1.5 x10 4 4.0 10 - 1 35 34 32 1.7 x10 2 3 50.26 234.00 10 0 166 159 145 7.8 x10 1 8.9 x10 1 4.2 x10 2 1.5 x10 4 2.8 10 - 1 21 23 16 1.0 x10 2 3 1 50.05 230.00 10 0 168 156 18 5 8.5 x10 1 1.0 x10 2 4.8 x10 2 6.9 x10 2 69.7 10 - 1 23 27 24 1.2 x10 2 2 50.06 231.00 10 0 35 28 38 9.0 x10 1 1.3 x10 2 6.0 x10 2 6.9 x10 2 87.0 10 - 1 165 193 183 1.7 x10 2 3 50.59 232.00 10 0 226 231 231 1.6 x10 2 1.4 x10 2 6.2 x10 2 6.8 x 10 2 91.1 10 - 1 22 39 32 1.2 x10 2 Avg NA*** NA*** NA*** NA*** NA*** NA*** NA*** NA*** 9.0x10 1 4.2x10 2 5.3x10 3 39.0 50 4.2 Yuma Irrigation Water Quality Yuma irrigation water sampling site descriptions as well as respective temperature, pH, turbidity, and conductivity measurements are prov ided in Table 8 . Overall , water conditions remained relatively stable betwe en all sampling sites. Tables 9 and 10 present the Yuma irrigation water results for enterococci and E. coli indicator bacteria , respectively. The average enterococci concentration for all six samples wa s 5.2 MPN/100mL w ith a standard deviation of 1.9 MPN/100 mL. Average E. coli and total coliform levels for all s ix samples were 151.6 MPN/100mL and 2.6 MPN/100mL, wit h a standard deviation of 109.6 and 1.4 respectively. Table 8 . Yuma irrigation water sample location descriptions and conditions Sample ID* Location description Temperature ( O C) pH Turbidity (NTU) Conductivity YW1 Opening of main canal from Imperial Dam. Carries Colorado River water into the Yuma Valley. 11.3 8.41 3.4 1134 YW 2 Yuma main canal at Picacho Road. Water has been carried through multiple agricultural areas. 11.5 8.44 3.6 1136 YW3 West main canal at West 2nd street and Ave B. Canal water previously suspected of septic tank contamination. 12.5 8.50 2.0 1137 YW4 Yuma main canal at West 1st street and Ave A, residential area. 13.2 8.45 1.6 1113 YW5 Yuma East Canal at Co. 18th street and Ave D, residential area. 11.8 8.41 2.3 1129 YW6 Yuma East Canal at 14th street and Ave B, near palm tree and agricultural fie lds. 12.3 8.44 1.8 1147 * Yuma irrigation water (YW) sample locations 1, 2, 3, 4, 5, and 6 51 * Yuma irrigation water (YW) sample locations 1, 2, 3, 4, 5, and 6 **Standard deviation = 1.91 ***NA: not applicable Table 10. Yuma irrigation water total coliform and E. coli levels Sample ID* Dilution Sample v olume (mL) # Large wells that are yellow # Small wells that are yellow # Large wells that fluoresce blue # Small wells that fluoresce blue Coliform MPN/100 mL 95% confidence limit ** E.coli MPN/100mL 95% confidence limit ** Lower Upper Lower Upper YW1 10 0 100 47 12 4 0 172.3 119.5 242.2 4.1 1.7 9.5 YW2 10 0 100 49 21 3 0 365.4 231.9 555.5 3.1 0.7 8.9 YW3 10 0 100 38 11 4 0 91 .0 66.6 121 .0 4.1 1.7 9.5 YW4 10 0 100 43 3 2 0 96 .0 68.5 132.1 2 .0 0.3 7.1 YW5 10 0 100 40 4 1 0 83.3 59.4 114.6 1 .0 0.1 5.5 YW6 1 0 0 100 41 9 1 0 101.4 74.3 136.1 1 .0 0.1 5.5 Average** NA*** NA*** 43 10 2.5 0 151.6 49.1 261.2 2.6 1.1 4.0 * Yuma irrigation water (YW) sample locations 1, 2, 3, 4, 5, and 6 ** Coliform standard deviation = 109.62; E.coli standard deviation = 1.43 *** NA: not applicable Table 9. Yuma irrigation water enterococci levels Sample ID* Dilution Volume (mL) # Large wells that fluoresce blue # Small wells that fluoresce blue Enterococ ci MPN/100mL 95% confidence limit ** Lower Upper YW1 10 0 100 5 3 8.4 3.7 15.3 YW2 10 0 100 3 2 5.1 1.7 10.6 YW3 10 0 100 4 0 4.1 1.7 9.5 YW4 10 0 100 6 0 6.3 2.9 13.7 YW5 10 0 100 4 0 4.1 1.7 9.5 YW6 10 0 100 3 0 3.1 0.7 8.9 Average** NA*** NA*** 4.2 2.5 5.2 3.3 7.1 52 4.3 Irrigation Water Metagenomic St atistics T able 11 shows the number of sequence reads following quality trimming, N50 statistic, and assembled contiguous sequence (contig) information f or each Yuma irrigation water sample . The i rrigation water virome resulted in 21.2 to 66.0 million sequence reads following quality trimming and 18.5 to 127.8 thousand assembled contigs larger than 200 bp. Average total assembly length for all samples was 36.2 mi llion base pairs with a maximum con tig size range of 12,172 to 61,978 bp. Table 11. Yuma irrigation water sequence statistics following trimming and a ssembly Sample * # Sequence reads following trimming # Contigs (> 200bp) Total assembly length (M bp) Max contig size (bp) N50* * YW1 65, 987,608 127,880 83.4 21,123 717 YW2 53,935,130 22,383 14.3 12,651 680 YW3 32,092,020 18,536 10.6 12,172 573 YW4 41,080,774 27,580 16.6 14,587 627 YW5 21,258,660 77,348 52.9 61,978 761 YW6 24,880,404 59,759 39.5 19,794 730 Average 39,872,433 55,581 36 23,718 681 *Yuma irrigation water (YW) sample locations 1, 2, 3, 4, 5, and 6 * * N50: genome assembly contig statistic 4.4 Irrigation Water V irome Table 12 and Figures 1 - 6 irrigation water viro me. For all six samples, the majority of contigs could not be annotated against the NCBI RefSeq viral database . The percentage of c ontigs with no hits ranged 64.5 - 84.5 % while the percentage o f those not assigned ranged 4.3 - 7.4 % (Figures 1 - 6). Of the assig ned viral sequence, the average majority ( 69.6 %) of contigs were assigned as dsDNA viruses while no more than 0.2 % shared sequence similarities with dsRNA viruses (Fig ures 1 - 6 ). Single - 53 stranded DNA and ssRNA viral sequences composed 5.3 - 31.3 and 1.7 - 12.3 % of the irrigation water virome respectively. Table 12. Distribution of contigs larger than 200 bp for Yuma irrigation water virome Genome * Samples ** YW1 # contigs YW 1 (%) YW2 # contigs YW 2 (%) YW3 # contigs YW 3 (%) YW4 # contigs YW 4 (%) YW5 # co ntigs YW5 (%) YW6 # contigs YW 6 (%) Avg (%) No Hits 107 , 829 84.5 14 , 533 65.4 13 , 022 70.5 19 , 023 69.1 49 , 695 64.5 40 , 505 67.9 70.3 Not Assigned 5 , 491 4.3 1 , 477 6.7 1 , 144 6.2 1 , 791 6.5 5 , 695 7.4 4 , 145 6.9 6.3 Assigned 14 , 337 11.2 6 , 196 27.9 4 , 308 23.3 6 , 711 24.4 21 , 680 28.1 14 , 991 25.4 23.4 dsDNA 9 , 260 64.6 2 , 593 41.8 3 , 294 76.5 5 , 498 81.9 14 , 938 68.9 12 , 597 84.0 69.6 dsRNA 21 0.1 13 0.2 2 0.0 3 0.0 1 0.0 9 0.1 0.1 Retro - transcribing 6 0.0 0 0.0 3 0.1 0 0.0 2 0.0 3 0.0 0.0 Satellites 34 0.2 27 0.4 11 0.3 11 0.2 29 0.1 17 0.1 0.2 ssDNA 2 , 381 16.6 1 , 936 31.2 444 10.3 359 5.3 3 , 618 16.7 792 5.3 14.3 ssRNA 880 6.1 764 12.3 175 4.1 240 3.6 854 3.9 253 1.7 5.3 U nclassified 1 , 755 12.2 863 13.9 379 8.8 600 8.9 2 , 238 10.3 1 , 320 8.8 10.5 *Contigs assigned to in the * *Yuma irrigation water (YW) sample locations 1, 2, 3, 4, 5, and 6 54 Fi Not Assigned, 5491 4.3% No Hits, 107829, 84.5% dsDNA, 9260 64.6% dsRNA, 21, 0.2% Retro - transcribing, 6, 0.0% Satellites,34, 0.2% ssDNA, 2381, 16.6% ssRNA, 880, 6.1% unclassified , 1755, 12.2% Assigned, 14337, 11.2% Figure 1. Distribution of contigs larger than 200bp for Yuma irrigation water 1 virome Not Assigned, 1477 6.7% No Hits, 14533, 65.5% dsDNA, 2593, 41.9% dsRNA, 13, 0.2% Satellites, 27, 0.4% ssDNA, 1936, 31.3% ssRNA, 764, 12.3% unclassified, 863, 13.9% Assigned, 6196, 27.9% Figure 2. Distribution of contigs larger than 200bp for Yuma irrigation water 2 virome 55 Not Assigned, 1144, 6.2% No Hits, 13022, 70.5% dsDNA, 3294, 76.5% dsRNA, 2, 0.1% Retro - transcribing , 3, 0.1% Satellites , 11, 0.3% ssDNA, 444, 10.3% ssRNA, 175, 4.1% unclassified, 379, 8.8% Assigned, 4308 23.3% Figure 3. Distribution of contigs larger than 200bp for Yuma irrigation water 3 virome Not Assigned, 1791, 6.5% No Hits, 19023, 69.1% dsDNA, 5498, 81.9% dsRNA, 3, 0.0% Satellites, 11, 0.2 % ssDNA, 359, 5.4% ssRNA, 240, 3.6% unclassified, 600, 8.9% Assigned, 6711, 24.4% Figure 4. Distribution of contigs larger than 200bp for Yuma irrigation water 4 virome 56 Not Assigned, 5696, 7.4% No Hits, 49695, 64.5% dsDNA, 14938, 68.9% Retro - transcribing, 2, 0.0% Satellites, 29, 0.1% ssDNA, 3618, 16.7% ssRNA, 854, 3.9% unclassified, 2238, 10.3% Assigned, 21680, 28.1% Figure 5. Distribution of contigs larger than 200bp for Yuma irrigation water 5 virome Not Assigned, 4145, 6.9% No Hits, 40505, 67.9% dsDNA, 12597, 84.0% dsRNA, 9, 0.1% Retro - transcribing, 3, 0.0% Satellites, 17, 0.1% ssDNA, 792, 5.3% ssRNA, 253, 1.7% unclassified, 1320, 8.8% Assigned, 14991, 25.4% Figure 6. Distribution of contigs larger than 200bp for Yuma irrigation water 6 virome 57 The distribution of viral host species for each irrigation water virome is shown in Table 13 and Figure 7 . Contigs assigned to viral orders and families were classified as algae, animals (vertebr ates only, invertebrates only, vertebrate and invertebrates), animals and plants, archaea, bacteria, or as other (amoeba, protozoa, and fungi). Bacteriophage dominated in all 6 irriga tion water samp les, comprising on average 78.7 % (range 57.2 - 87.7 %) of th e irrigation water virome. Of the remaining sing le host categories, algae (4.8 %), pla nts (2.6 %), inverteb rate animals (2.7 %) , and vertebrate animals (0.7 %) comprised a far smaller average percentage of the irrigation water virome. The average percentage of viral sequences assigned to eukaryotic, multiple host categories including vertebrate and invertebrate animals, animals and plants, as well as eukaryotic amoeba, protozoa, and fung i (classified as other) was 4.8 %, 0.9 %, 4.8%, respectively. 58 Ta ble 13. Yuma irrigation water viral host distribution Sample* Host YW1 # contigs YW1 (%) YW2 # contigs YW2 (%) YW3 # contigs YW3 (%) YW4 # contigs YW4 (%) YW5 # contigs YW5 (%) YW6 # contigs YW6 (%) Avg (%) Algae 1 , 004 9.9 134 3.0 135 4.0 219 4.1 566 3.3 523 4.4 4.8 Animals (invertebrate) 460 4.5 217 4.8 77 2.3 92 1.7 310 1.8 122 1.0 2.7 Animals (vertebrate and invertebrate) 1 , 017 10.0 384 8.6 121 3.6 107 2.0 433 2.5 279 2.3 4.8 Animals (vertebrate) 170 1.7 46 1.0 11 0.3 15 0.3 62 0.4 38 0.3 0.7 An imals and Plants 125 1.2 99 2.2 16 0.5 34 0.6 113 0.7 30 0.3 0.9 Archaea 4 0.0 2 0.0 0 0.0 3 0.1 6 0.0 4 0.0 0.0 Bacteria 5 , 804 57.2 3 , 216 71.7 2 , 806 83.0 4 , 598 85.7 15 , 046 87.7 10 , 349 86.7 78.7 Other (amoeba, protozoa, or fungi) 1 , 203 11.9 82 1.8 147 4 .3 240 4.5 380 2.2 485 4.1 4.8 Plants 364 3.6 305 6.8 67 2.0 58 1.1 242 1.4 112 0.9 2.6 * Yuma irrigation water (YW) sample locations 1, 2, 3, 4, 5, and 6 59 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% YW1 YW2 YW3 YW4 YW5 YW6 Conti g Distribution Sample ID Figure 7. Viral host distribution in six Yuma irrgation water samples Plants Other (amoeba, protozoa, or fungi) Bacteria Archaea Animals and Plants Animals (vertebrate) Animals (vertebrate and invertebrate) Animals (invertebrate) Algae 60 A total of 64 viral families were observed for all 6 irrigation water samples. The viral fam ily contig distribution can be viewed in a phylogenetic tree constructed for all six irrigation water sample s using MEGAN software (Figure 8 ). The number of assigned contigs in this figure is represented by circle size and colors show the proporti on belon ging to each sample . A large proportion of viruses for all 6 samples were assigned to Myoviridae, Podoviridae, Siphoviridae, and Microviridae bacteriophage families. The distribution of viral contigs associated with families composing .0% across the si x irrigation water viromes w as further analyzed (Table 14 and Figure 9) . O .0 % of the irrigation water virome, a majority were assigned to Myoviridae (28.6 %), Podoviridae ( 16.5 %) and Siphoviridae ( 20.5 %) dsDNA bacteriophage families of the Cau dovirales order , however the Microviridae (15.7 %) bacteriophage family was prevalent in the irrigation water virome as well. Other viral families that were well represented in irrigation water included those infecting animals ( Circovirid ae , Dicistroviridae , and Poxviridae ) , algae ( P hycodnaviridae ), and amoeba ( Mimiviridae ). The contig distribution of phage hos t species is provided in Table 15 . Analysis of the bacteriophage host species revealed a total of 88 bacterial hosts. Host speci .0 % across the 6 irrigation water viromes included Pelagibacter (11.9 - 16.5 %), Cellulophaga (10.3 - 13.6 %), Synechococcus (12.7 - 13.2 %), Bdellovibrio (0.9 - 8. 3%), Bacillus (3.1 - 4.4 %), Pu niceispirillium (2.6 - 3.8 %), Mycobacterium (2.0 - 3.7 %), an d Prochloro coccus (2.8 - 3.2 %). Phage infecting potential bacterial foodborne pathogens including Campylobacter (0.7%), Enterococcus (0.3%), Escherichia (1.9%), Listeria (0.1%), Salmonella (2.1%), and Shigella (0.0%) on average composed a much lower percent age of the bacterial host species. 61 Figure 8 . Yuma irrigation water viral family distribution 62 Table 14. Irrigation water viral family distribution representing greater than 3.0% of the virome Sample* Family YW1 # contigs YW1 ( %) YW2 # contigs YW2 (%) YW3 # contigs YW3 (%) YW4 # contigs YW4 (%) YW5 # contigs YW5 (%) YW6 # contigs YW6 (%) Avg (%) Circoviridae 350 4.4 304 8.8 75 2.8 58 1.4 257 1.9 147 1.6 3.5 Dicistroviridae 198 2.5 188 5.4 60 2.3 56 1.3 234 1.7 65 0.7 2.3 Micr oviridae 1 , 471 18.5 1 , 174 33.9 255 9.6 233 5.6 2 , 900 21.6 440 4.8 15.7 Mimiviridae 1 , 080 13.6 67 1.9 136 5.1 225 5.4 354 2.6 458 5.0 5.6 Myoviridae 1 , 460 18.4 683 19.7 894 33.8 1 , 498 36.1 3 , 595 26.7 3 , 400 36.8 28.6 Phycodnaviridae 997 12.6 131 3.8 133 5 .0 216 5.2 560 4.2 517 5.6 6.1 Podoviridae 770 9.7 416 12.0 473 17.9 842 20.3 2 , 488 18.5 1 , 820 19.7 16.3 Poxviridae 503 6.3 7 0.2 12 0.5 16 0.4 45 0.3 32 0.3 1.3 Siphoviridae 1 , 111 14.0 489 14.1 607 22.9 1 , 008 24.3 3 , 019 22.4 2 , 352 25.5 20.5 * Yuma irr igation water (YW) sample locations 1, 2, 3, 4, 5, and 6 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% YW1 YW2 YW3 YW4 YW5 YW6 Contig Distribution Sample Figure 9. Irrigation water viral family distribution representing greater than 3.0% of the virome Siphoviridae Poxviridae Podoviridae Phycodnaviridae Myoviridae Mimiviridae Microviridae Dicistroviridae Circoviridae 63 Table 15 . Bacteriophage host distribution of Yuma irrigation water samples Host YW1 (%) YW2 (%) YW3 (%) YW4 (%) YW5 (%) YW6 (%) Average (%) Acinetobacter 0.3 0.3 0.5 0.5 0.6 0.6 0.5 Actinomyces 0.1 0.1 0.1 0.3 0.0 0.0 0.1 Actinoplanes 0.2 0.5 0.3 0.0 0.3 0.2 0.2 Aeromonas 0.2 0.2 0.4 0.3 0.4 0.3 0.3 Aggregatibacter 0.1 0.2 0.1 0.2 0.1 0.1 0.1 Agrobacterium 0.2 0.0 0.0 0.1 0.1 0.1 0.1 Altermonas 0.6 0.4 0.2 0.5 0.5 0.5 0.4 Arthr obacter 0.1 0.2 0.0 0.2 0.2 0.2 0.1 Azospirillum 0.2 0.4 0.4 0.2 0.4 0.2 0.3 Bacillus 3.7 3.7 4.4 4.2 3.1 3.5 3.8 Bacteroids 0.1 0.0 0.1 0.0 0.1 0.1 0.1 Bdellovibrio 5.9 8.3 1.3 1.1 6.0 0.9 3.9 Brochothrix 0.1 0.1 0.1 0.1 0.0 0.0 0.1 Burkholderia 1.6 1.5 1.4 1.1 1.6 1.7 1.5 Candidatus 0.1 0.0 0.0 0.1 0.0 0.0 0.0 Campylobacter 0.9 0.9 0.7 0.6 0.6 0.4 0.7 Caulobacter 1.2 0.4 0.7 0.6 0.9 0.7 0.8 Cellulophaga 11.2 13.6 10.3 13.2 11.4 10.5 11.7 Chlamydia 0.7 0.8 0.1 0.1 0.5 0.1 0.4 Clavibacter 0.1 0. 0 0.1 0.0 0.3 0.2 0.1 Clostridium 0.7 0.7 1.2 0.7 0.7 0.8 0.8 Colwellia 0.1 0.2 0.3 0.1 0.1 0.1 0.1 Corynebacterium 0.2 0.4 0.3 0.2 0.2 0.2 0.2 Croceibacter 0.3 0.3 0.3 0.3 0.3 0.4 0.3 Cronobacter 2.4 2.2 2.1 2.3 1.7 2.1 2.1 Deftia 0.3 0.3 0.2 0.3 0. 2 0.3 0.2 Dickeya 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Edwardsiella 0.1 0.0 0.1 0.1 0.1 0.2 0.1 Enterobacter 0.1 0.3 0.2 0.1 0.1 0.1 0.1 Enterococcus 0.3 0.4 0.1 0.2 0.5 0.2 0.3 Erwinia 0.6 0.1 0.2 0.2 0.3 0.2 0.2 Escherichia 2.5 1.4 1.6 2.0 2.0 1.8 1.9 Flav obacterium 1.2 0.9 1.3 1.6 1.4 1.4 1.3 Gordonia 0.1 0.0 0.0 0.0 0.1 0.0 0.0 Haemophilus 0.0 0.1 0.0 0.0 0.0 0.1 0.0 Iodobacter 0.1 0.1 0.2 0.1 0.1 0.1 0.1 Klebsiella 0.9 0.4 0.3 0.6 0.5 0.6 0.5 Lactobacillus 0.7 0.6 0.6 0.5 0.9 0.8 0.7 Lactococcus 1. 7 0.8 1.5 1.0 0.8 1.1 1.2 Listeria 0.1 0.1 0.1 0.0 0.3 0.2 0.1 Listonella 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Methanobacterium 0.1 0.0 0.1 0.0 0.0 0.0 0.0 Methanothermobacter 0.0 0.0 0.0 0.1 0.0 0.0 0.0 Microbacterium 0.5 0.5 0.4 0.4 0.2 0.4 0.4 Mycobacteriu m 3.7 2.1 3.6 2.5 2.0 2.2 2.7 Myxococcus 1.0 1.3 0.7 0.6 1.6 0.8 1.0 64 * Viral contigs assigned as phage but are not assigned or classified into single host category ** Bolded numbers represent % of any individual Yuma irrigation water virome Natrialba 0.1 0.0 0.3 0.2 0.2 0.3 0.2 Nocardia 0.1 0.1 0.1 0.1 0.0 0.1 0.1 Paenibacillus 0.0 0.1 0.2 0.1 0.1 0.1 0.1 Pantoea 0.1 0.1 0.3 0.2 0.1 0.1 0.1 Pasteurel la 0.0 0.0 0.1 0.0 0.0 0.0 0.0 Pectobacterium 0.1 0.1 0.3 0.2 0.2 0.1 0.2 Pelagibacter 11.9 13.2 16.5 14.9 12.5 14.7 13.9 Phormidium 0.0 0.0 0.1 0.0 0.1 0.0 0.0 Planktothrix 1.9 1.8 2.2 3.1 1.6 3.0 2.3 Prochlorococcus 3.5 2.8 3.1 3.6 3.3 3.9 3.4 Prop ionibacterium 0.0 0.1 0.1 0.1 0.0 0.0 0.1 Pseudoalteromonas 0.4 0.3 0.3 0.3 0.3 0.4 0.3 Pseudomonas 2.5 2.6 3.0 3.2 2.8 2.9 2.8 Psychrobacter 0.4 0.6 0.9 0.5 0.7 0.6 0.6 Puniceispirillum 3.8 2.6 3.7 3.5 4.4 4.7 3.8 Ralstonia 0.7 0.3 1.2 1.0 1.4 1.3 1. 0 Rhizobium 0.5 1.0 1.1 1.1 0.8 1.0 0.9 Rhodobacter 0.0 0.1 0.1 0.1 0.0 0.0 0.1 Rhodococcus 1.0 0.7 0.5 0.7 0.8 0.8 0.8 Rhodothermus 2.2 3.4 2.5 2.2 2.0 2.0 2.4 Riemerella 0.3 0.5 0.4 0.4 0.6 0.4 0.4 Roseobacter 0.2 0.5 0.5 0.5 0.5 0.4 0.4 Salinivib rio 0.1 0.1 0.3 0.2 0.2 0.1 0.2 Salmonella 1.7 2.1 2.5 2.7 1.8 1.8 2.1 Serratia 0.1 0.0 0.0 0.2 0.0 0.0 0.1 Shigella 0.1 0.0 0.0 0.1 0.1 0.0 0.0 Sodalis 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Sphingomonas 1.1 0.3 1.2 1.4 0.5 0.7 0.9 Spiroplasma 0.7 0.5 0.1 0.0 0.2 0.0 0.3 Staphylococcus 0.4 0.5 0.1 0.3 0.1 0.1 0.2 Stenotrophomonas 0.1 0.0 0.2 0.2 0.1 0.1 0.1 Streptococcus 0.7 0.5 0.5 0.6 0.7 0.9 0.7 Streptomyces 0.8 0.5 0.5 0.6 0.5 0.9 0.6 Synechococcus 13.2 12.7 14.4 15.6 15.3 19.0 15.0 Thalassomonas 0.2 0.5 0.5 0.7 0.7 0.3 0.5 Thermoanaerobacterium 0.7 0.3 0.4 0.5 1.1 0.5 0.6 Thermus 0.1 0.0 0.0 0.1 0.0 0.0 0.0 Tsukamurella 0.0 0.0 0.1 0.2 0.1 0.1 0.1 Vibrio 1.6 2.2 2.5 2.4 2.6 2.4 2.3 Xanthomonas 0.4 0.5 0.5 0.2 0.2 0.4 0.4 Xylella 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Yersinia 0.3 0.5 0.9 0.7 0.4 0.7 0.6 Unknown host genus* 3.5 4.0 1.6 1.4 2.0 1.1 2.3 65 For all six irrigation water samples, species of interest to the food industry belonging to 30 virus families were grouped into 5 categories based on the hosts infected including agricultural insect pests, crops, commercial fish or shrimp (fresh and salt water), livestock, and human pathogens. A summary of virus families, hosts, possible species of interest to the food industry, and % contig identities are presented in Table 16 . Virus families that infect agricultural insect pests included Alphatetraviridae , Ascoviridae , Baculoviridae , Dicistroviridae , Iflaviridae, Nudiviridae , Parvoviridae , and Poxviridae . Recovered virus hosts including the beet armyworm, corn earworm, green peach aphid, cabbage looper, and alfalfa looper are all known agricultural insect pests that damage lettuce. Many popular plant and crop viruses belonged to single - stranded RNA families including Benyviridae , Bromoviridae , Closteroviridae , Secoviridae, Tombusviridae , Tymoviridae, and Virgaviridae. In addition, irrigation water contained numerous crop single - stranded DNA viral pathogens belonging to the Geminiviridae and Nanoviridae families. Viruses infecting commercial fish and shrimp were also recovered, represented by the Circovir idae , Hepeviridae , Iridoviridae , Nimaviridae , Roniviridae, and Totiviridae virus families. Relevant livestock pathogens infecting pigs ( African swine fever virus , pseudorabies , swinepox virus, and porcine circovirus (type 1/2a), parvovirus , astrovirus , an d teschovirus ), cow ( cowpox , bovine papular stomatitis virus , bovine hungarovirus 1, Enterovirus F ), and poultry ( fowlpox and turkey hepatitis virus 2993D ) belonged to the Asfarviridae , Astroviridae , Circoviridae , Herpesviridae , Parvoviridae , Picornavirida e , and Poxviridae virus families. It is important to note tha t the Iridoviridae, Poxviridae, Circoviridae, and Parvoviridae viral families infect a wide variety of both vertebrate and invertebrate animal hosts; therefore host infection will vary with vira l species. H uman viral pathogens including picobirnavirus , coronavirus , parechovirus , and rhinovirus as well as hepatitis A and E virus 66 which are commonly transmitted through fresh produce were identified in the irrigation samples as well. A majority of virus species had a low percent contig identity with the exceptions of Spodoptera exigua iflavirus 1 (93 - 100%), Tobacco necrosis virus D (93 - 97%), and Cucumber green mottle mosaic virus (96%). Table 16 . Viral families and species of interest identified in irrigation water Species relevance Species of interest Virus f amily Species contig identity range (%) Agricultural insect pest Helicoverpa armigera stunt virus (Old World cotton bollworm) Alphatetraviridae 21 Spodoptera frugiperda ascovirus 1a (fall ar myworm) Ascoviridae 29 Trichoplusia ni ascovirus 2c (cabbage looper) 22 - 44 Heliothis virescens ascovirus 3a (tobacco budworm) 18 Agrotis segetum granulovirus (turnip moth) Baculoviridae 22 Anticarsia gemmatalis nucleopolyhedrovirus (velvetbean ca terpillar) 29 Autographa californica nucleopolyhedrovirus (alfalfa looper) 29 Clanis bilineata nucleopolyhedrovirus (soybean pest) 33 - 36 Cryptophlebia leucotreta granulovirus (false codling moth) 22 - 27 Cydia pomonella granulovirus (codling moth ) 40 Phthorimaea operculella granulovirus (potato tuber moth) 21 - 31 Plutella xylostella granulovirus (diamondback moth) 45 Pseudaletia unipuncta granulovirus (white - speck moth) 29 Pieris rapae granulovirus (cabbage butterfly) 46 Spodoptera l itura nucleopolyhedrovirus (oriental leafworm moth) 29 Trichoplusia ni single nucleopolyhedrovirus (cabbage looper) 23 Himetobi P virus (Small brown planthopper) Dicistroviridae 21 - 46 Homalodisca coagulata virus - 1 (glassy - winged sharpshooter) 21 - 6 8 Rhopalosiphum padi virus (bird cherry oat aphid) 19 - 98 Solenopsis invicta virus - 1 (imported fire ant) 23 - 45 Brevicoryne brassicae picorna - like virus (cabbage aphid) Iflaviridae 27 - 34 67 Table 16 ( ) Nilaparvata lugens honeydew virus - 2 (brown planthopper) 26 - 29 Spodoptera exigua iflavirus 1 (beet armyworm) 93 - 100 Helicoverpa zea nudivirus 2 (corn earworm) Nudiviridae 21 - 60 Helicoverpa armigera densovirus (Old World cotton bollworm) Parvoviridae * 26 - 40 Myzus persicae densovirus (green peach aphid) 28 Planococcus citri densovirus (citrius mealybug) 38 - 86 Adoxophyes honmai entomopoxvirus 'L' (smaller tea tortix) Poxviridae * 26 Amsacta moorei entomopoxvirus 'L' (tiger moth) 23 - 42 Choristoneura rosaceana entomopoxvirus 'L' (obli que banded leafroller) 26,29 Melanoplus sanguinipes entomopoxvirus (migratory grasshopper) 21 - 47 Mythimna separata entomopoxvirus 'L' (northern armyworm) 24 - 30 Crop virus Beet soil - borne mosaic virus Benyvirus (unassigned family) 23 - 32 Prune dwar f virus Bromoviridae 36 Grapevine rootstock stem lesion associated virus Closteroviridae 30 Persea americana endornavirus (avacado) Endornaviridae 25 Chickpea chlorosis Geminiviridae 24 Chickpea redleaf virus 24 - 32 Citrus chlorotic dwarf associa ted virus 35 - 42 Maize streak Reunion virus 24 Melon chlorotic mosaic virus 34 Papaya leaf curl China virus 38 Pepper golden mosaic virus 34 Tomato mild mosaic virus 49 Tomato yellow leaf curl virus 32 Soybean chlorotic spot virus 28 Watermelon chlorotic stunt virus 31 Wheat dwarf virus 27 - 37 Abaca bunchy top virus Nanoviridae 41 Banana bunchy top virus 39 - 43 Pea necrotic yellow dwarf virus 33 Cassava virus C Ourmiavirus (unassigned family) 23 - 36 Epirus cherry virus 3 0 - 46 Ourmia melon virus 21 - 43 Blackcurrant reversion virus Secoviridae 32 68 Table 16 ( ) Maize chlorotic dwarf virus 35 Satsuma dwarf virus 41 Strawberry latent ringspot virus 33 Rice tungro spherical virus 25 - 34 Tomato ringspot viru s 25 Southern bean mosaic virus Sobemovirus (unassigned family) 79 Rice grassy stunt virus Tenuvirus (unassigned family) 32 Beet black scorch virus Tombusviridae 24 Cucumber bulgarian virus 50 - 72 Cucumber leaf spot virus 27 - 54 Cucumber necro sis virus 80 Maize chlorotic mottle virus 34 - 46 Maize white line mosaic virus 29 Melon necrotic spot virus 57,59 Oat chlorotic stunt virus 32 - 71 Olive mild mosaic virus 87 Soybean yellow mottle mosaic virus 34 Tobacco necrosis virus D 93,97 Grapevine fleck virus Tymoviridae 28 Carrot mottle mimic virus Umbravirus (unassigned family) 33 Cucumber green mottle mosaic virus Virgaviridae 96 Fish and shrimp pathogens Penaeus monodon circovirus VN11 Circoviridae * 29 - 54 Cutthroat trou t virus Hepeviridae 19 - 34 Lymphocystis disease virus 1 (European flounder and plaice) Iridoviridae * 18 Lymphocystis disease virus - isolate China (flounder) 25 - 50 Infectious spleen and kidney necrosis virus (fish) 20 - 48 Singapore grouper iridovir us 20 - 42 Shrimp white spot syndrome virus Nimaviridae 26 - 57 Gill - associated virus (black tiger prawn) Roniviridae 27 - 30 Penaeid shrimp infectious myonecrosis virus Totiviridae 40 Human pathogen Human coronavirus HKU1 Coronaviridae 39 Hepatitis E virus Hepeviridae 23 - 34 Human Picobirnavirus Picobirnaviridae 26 - 34 Hepatitis A virus Picornaviridae 26 - 34 Human parechovirus 26 69 Table 16 ( ) Rhinovirus C 22 Livestock pathogen African swine fever virus Asfarviridae 27 - 30 Porcine astrovi rus 3 Astroviridae 24 Porcine circovirus type 1/2a Circoviridae * 31 - 47 Suid herpesvirus 1 (pseudorabies) Herpesviridae 27 Porcine parvovirus Parvoviridae * 32,43 Bovine hungarovirus 1 Picornaviridae 50 Enterovirus F (bovine) 34 Porcine teschovi rus 36 Turkey hepatitis virus 2993D 36 Bovine papular stomatitis virus Poxviridae * 39 Cowpox virus 23 - 35 Fowlpox virus 25 - 37 Swinepox virus 24 - 30 *Viral families with wide host range , host depends on specific viral species 4.5 Yuma Lettuce Virome Statistics Table 17 shows the number of sequence reads following quality trimming, N50 statistic, and assembled contiguous sequence (contig) information f or Yuma lettuce samples . The romaine lettuce virome resulted in ~14.3 to 40.3 million seque nce reads following quality trimming and 887 to 3,491 assembled contig s larger than 200 bp. T otal assembly length for all romaine lettuce samples was 1.2 mi llion base pairs with a maximum contig size range of 5,875 to 59,171 bp. The iceberg lettuce virom e resulted in ~15.5 to 45.2 million sequence reads following quality trimming and 703 to 5,577 assembled contigs larger than 200 bp. Average total assembly length for all iceberg lettuce samples was 1.3 mi llion base pairs with a maximum contig size range of 2,623 to 58,547 bp. 70 Table 17. Yuma lettuce sequence s tatistic s following trimming and a ssembly Lettuce Type Sample ID Sample Description # Sequence reads following trimming # contigs (>200bp) Total assembly l ength (bp) Max contig size (bp) N50 Romai ne YL1 Worker harvest 22,192,992 1,620 962,920 8,632 580 YL2 Worker harvest 17,200,830 2,275 1,334,015 10,091 582 YL3 Worker harvest 16,977,540 1,086 722,002 20,312 652 YL4 Control 29,424,106 2,251 1,247,443 6,476 543 YL5 Control 35,041,916 1,663 9 64,267 12,624 553 YL6 Control 29,026,710 2,002 1,035,832 7,519 492 Iceberg YL7 Control 29,742,342 2,720 1,330,578 12,556 474 YL8 Control 25,153,498 2,549 1,176,162 5,526 440 YL9 Control 26,246,806 1,981 1,075,789 5,135 526 YL10 Worker harvest 21,4 06,636 1,262 841,841 13,516 694 YL11 Worker harvest 16,549,704 3,475 1,854,371 8,344 537 YL12 Worker harvest 20,684,240 784 511,741 8,381 662 YL13 Worker harvest 18,585,596 3,315 2,236,049 19,370 703 YL14 Worker harvest 20,012,550 1,292 1,182,495 5 8,547 1,092 YL15 Worker harvest 23,122,222 1,194 911,149 11,376 825 YL16 Worker harvest 24,312,410 703 596,082 11,488 1,064 YL17 Worker harvest 25,204,382 2,435 1,694,477 11,427 717 YL18 Control 29,990,912 2,201 1,255,819 37,446 570 YL19 Control 28,146,472 1,699 897,536 5,206 528 YL20 Control 45,206,036 2,980 1,689,182 38,197 566 YL21 Control 38,761,924 3,312 2,022,747 9,904 610 YL22 Control 36,458,692 2,072 1,093,235 8,085 527 YL23 After worker 30 min break 15,534,726 5,577 3,602,163 21,5 40 692 YL24 After worker 30 min break 17,791,284 1,641 987,166 4,163 639 YL25 After worker 30 min break 17,898,636 713 605,017 12,873 974 71 YL26 After worker 30 min break 16,128,066 845 477,225 2,623 583 YL27 After worker 30 min break 18,140,228 2,644 1,901,198 13,928 791 Romaine YL28 Control 37,243,138 1,446 941,058 27,199 680 YL29 Control 34,020,132 1,223 873,414 59,171 781 YL30 Control 40,302,750 1,229 841,510 16,353 728 YL31 Control 37,703,770 1,637 1,062,459 41,708 675 YL32 Control 39,795,730 1,767 1,670,543 58,805 1,307 YL33 Worker harvest 14,261,170 1,562 1,321,166 16,909 938 YL34 After chop and wash 15,191,256 887 776,029 12,523 996 YL35 After chop and wash 18,912,828 3,491 2,411,520 20,547 707 YL36 After c hop and wash 16,904,810 3,342 2,384,372 21,488 795 YL37 Mixed salad 20,698,078 992 772,902 12,779 820 YL38 Mixed salad 19,025,938 1,168 836,064 5,875 788 YL39 Mixed salad 18,618,602 925 671,693 6,841 765 YL40 Mixed salad 17,769,542 1,766 1,280,403 12,422 759 YL41 Mixed salad 18,109,612 3,145 2,085,159 15,874 692 YL42 Worker harvest 20,578,110 2,078 1,418,464 18,786 686 4.6 Yuma Lettuce Virome To analyze the iceberg (N=21) and romaine (N=21) lettuce virome, the viral contigs belonging to mult iple samples collected from the same stage of field production were combined into a single sample. The resulting categories include iceberg (N=8) and romaine (N=8) lettuce controls (harvested by research crew), iceberg (N=8) and romaine (N=5) lettuce work er harvest, iceberg harvested post worker break (N=5), romaine chop and wash (N=3), and romaine mixed salad (N=5). The viral genome distribution for each of these categories is shown in Table 18 and Figures 10 - 16 . The majority of contigs belonging to bot h iceberg and romaine lettuce could not be annotated against the NCBI RefSeq viral database . 72 Table 18. Distribution of contigs larger than 200 bp for iceberg and romaine lettuce virome Genome * Sample Type ** IC # contigs IC (%) IWH # contigs IWH (%) IPWB # contigs IPWB (%) I ce Avg (%) RC # contigs RC (%) RWH # contigs RWH (%) RCW # contigs RCW (%) RMS # contigs RMS (%) Rom Avg (%) No Hits 17 , 004 87.1 12 , 563 86.9 10 , 288 90.1 88.0 10 , 611 80.3 7 , 298 84.7 6 , 962 90.2 6 , 884 86.1 85.3 Not Assigned 388 2.0 252 1.7 179 1.6 1.8 239 1.8 148 1.7 96 1.2 116 1.5 1.6 Assigned 2 , 122 10.9 1 , 644 11.4 953 8.4 10.2 2 , 366 17.9 1 , 175 13.6 662 8.6 994 12.4 13.1 dsDNA 1 , 215 57.3 613 37.3 397 41.7 45.4 1 , 370 57.9 385 32.8 124 18.7 386 38.8 37.1 dsRNA 59 2.8 141 8.6 51 5.4 5.6 250 10.6 237 20.2 98 14.8 160 16.1 15.4 Retro - transcribing 371 17.5 600 36.5 369 38.7 30.9 275 11.6 298 25.4 329 49.7 282 28.4 28.8 Satellites 0 0.0 0 0.0 0 0.0 0.0 2 0.1 0 0.0 0 0.0 0 0.0 0.0 ssDNA 132 6.2 53 3.2 29 3.0 4.2 59 2.5 38 3.2 18 2.7 22 2.2 2.6 ssRNA 178 8.4 140 8.5 55 5.8 7.6 193 8.2 134 11.4 55 8.3 80 8.1 9.0 Unclassified 167 7.9 97 5.9 52 5.5 6.4 217 9.2 83 7.1 38 5.7 64 6.4 7.1 * Contigs assigned to viral taxa but did not meet the selected MEGAN parameters are * * Iceberg control (IC), iceberg worker harvested (IWH), iceberg post worker break (IPWB), iceberg average (ice avg), romaine control (RC), romai ne worker harvested (RWH), romaine chop and wash (RCW), romaine mixed salad (RMS) , romaine average (rom avg). 73 No Hits, 17004, 87.1% Not Assigned, 388, 2.0% dsDNA, 1215 57.3% dsRNA, 59, 2.8% Retro - transcribing, 371, 17.5% ssDNA, 132, 6.2% ssRNA, 178, 8.4% Unclassified, 167, 7.9% Assigned , 2122, 10.9% Figure 10. Iceberg lettuce - control genome distribution N=8 No Hits, 12563, 86.9% Not Assigned, 252, 1.7% dsDNA, 613, 37.3% dsRNA, 141, 8.6% Retro - transcribing, 600, 36.5% ssDNA, 53, 3.2% ssRNA, 140, 8.5% Unclassified, 97, 5.9% Assigned, 1644, 11.4% Figure 11. Iceberg lettuce - worker harvest genome distribution N=8 74 No Hits, 10288, 90.1% Not Assigned, 179, 1.6% dsDNA, 397, 41.7% dsRNA, 51, 5.4% Retro - transcribing, 369, 38.7% ssDNA, 29, 3.0% ssRNA, 55, 5.8% Unclassified, 52, 5.5% Assigned, 953, 8.4% Figure 12. Iceberg lettuce - post worker break genome distribtuion N=5 No Hits, 10611, 80.3% Not Assigned, 239, 1.8% dsDNA, 1370, 57.9% dsRNA, 250, 10.6% Retro - transcribing, 275, 11.6% Satellites, 2, 0.1% ssDNA, 59, 2.5% ssRNA, 193, 8.2% Unclassified, 217, 9.2% Assigned, 2366, 17.9% Figure 13. Romaine lettuce - control genome distribution N=8 75 No Hits, 7298, 84.7% Not Assigned, 148, 1.7% dsDNA, 385, 32.8% dsRNA, 237, 20.2% Retro - transcribing, 298, 25.4% ssDNA, 38, 3.2% ssRNA, 134, 11.4% Unclassified, 83, 7.1% Assigned, 1175, 13.6% Figure 14. Romaine lettuce - worker harvest genome distribution N=5 No Hits, 6962, 90.2% Not Assigned, 96, 1.2% dsDNA, 124, 18.7% dsRNA, 98, 14.8% Retro - transcribing, 329, 49.7% ssDNA, 18, 2.7% ssRNA, 55, 8.3% Unclassified, 38, 5.7% Assigned, 662, 8.6% Figure 15. Romaine lettuce - chop and wash genome distribution N=3 76 The average number of assigned contigs >200 bp was 10.2 % and 13.1 % for iceberg and romaine lettuce samples respectively. Of the assigned viral sequence s belonging to iceberg lettuce , the majority of contigs shared similarities with dsDNA (average of 45.4%) and retro - transcribing (ave rage of 30.9 %) v iruses. Single - stranded DNA, dsRNA, and ssRNA viral sequences composed on aver age 4.2%, 5.6%, and 7.6 % of the iceberg lettuce virome , respectively. Similar results were observed for romaine lettuce viral contigs, with dsDNA and retro - transcribing viruses representing an average of 37.1% and 28.8 % of the virome , respectively. The remaining romaine lettuce viral sequen ces were assigned to ssDNA (2.6%), dsRNA (15.4 %), and ssRNA (9.0 %) viruses. No Hits, 6884, 86.1% Not Assigned, 116, 1.5% dsDNA, 386, 38.8% dsRNA, 160, 16.1% Retro - transcribing, 282, 28.4% ssDNA, 22, 2.2% ssRNA, 80, 8.1% Unclassified, 64, 6.4% Assigned, 994, 12.4% Figure 16. Romaine lettuce - mixed salad genome distribution N=5 77 The distribution of viral contigs based on virus hosts is show n in Table 19 and Figure 17. Contigs assigned to viral orders and families were classified into the same host categories as Yuma irrigation water samples. For iceberg lettuce samples (control, worker harvest, and post worker break), the majority of viral contigs were associated with bacteriophage and plant hosts, r epresenting an average of 40.4% and 30.7 % of the iceberg virome respectively. Of the remaining host categories, algae, invertebrates, vertebrates, and multiple hosts (invertebrates and vertebra tes, animals and plants, and other) composed on average 3. 5 %, 1. 4%, 12.5%, and 11.6 % of the iceberg lettuce virome , respectively. Similar to iceberg lettuce samples, the majority of viral contigs belonging to romaine lettuce harvested by researchers (cont rol), workers, as well as sampled from mixed salad were associated with bacteria (average of 39.9 %) a nd plant hosts (average of 28.2 %). However for romaine lettuce sampled after chop and wash, the majority of contigs we re associated with plants (39.2%), v ertebrate animals (24. 5%), and hosts classified as %), with bact eriophage comprising only (11.2 %) of the virome (Table 19). Algae, invertebrates, vertebrates, and multiple hosts (invertebrates and vertebrates, animals and plants, an d other) c omposed on average 2.6%, 1.9%, 12. 8%, and 19.1 % of the romaine lettuce virome , respectively. 78 Table 19. Iceberg and romaine lettuce viral host distribution Host Sample Type * IC # contigs IC ( %) IWH # contigs IWH (%) IPWB # contigs IPWB (%) Ice Av g (%) RC # contigs RC (%) RWH # contigs RWH (%) RCW # contigs RCW (%) RMS # contigs RMS (%) Rom Avg (%) Algae 76 4.2 42 3.0 27 3.4 3.5 29 1.4 27 2.7 13 2.4 31 3.7 2.6 Animals (invertebrate) 27 1.5 21 1.5 9 1.1 1.4 43 2.1 26 2.6 9 1.7 10 1.2 1.9 Animals (vertebrate and invertebrate) 47 2.6 42 3.0 22 2.8 2.8 40 1.9 21 2.1 19 3.5 19 2.3 2.5 Animals (vertebrate) 108 6.0 196 13.9 140 17.6 12.5 70 3.4 107 10.8 133 24.5 105 12.7 12.8 Animals and plants 4 0.2 28 2.0 12 1.5 1.2 5 0.2 24 2.4 6 1.1 18 2.2 1.5 Bacteria 1 , 017 56.7 446 31.7 261 32.7 40.4 1 , 281 61.5 280 28.3 61 11.2 246 29.7 32.7 Other (ameoba, protozoa, or fungi) 117 6.5 126 8.9 57 7.2 7.5 216 10.4 179 18.1 90 16.5 128 15.5 15.1 Plants 398 22.2 508 36.1 269 33.8 30.7 399 19.2 324 32.8 213 39.2 271 32.7 31.0 * Iceberg control (IC), iceberg worker harvested (IWH), iceberg post worker break (IPWB), iceberg average (ice avg), romaine control (RC), romaine worker harvested (RWH), romaine chop and wash (RCW), romaine mixed salad (RMS) , romaine average (rom avg). 79 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Iceberg- Control Iceberg- Worker Harvest Iceberg- Post Worker Break Romaine- Control Romaine- Worker Harvest Romaine- Chop and Wash Romaine- Mixed Salad Contig Distribution Sample Type Figure 17. Yuma lettuce viral host distribution Plants Other (ameoba, protozoa, or fungi) Bacteria Animals and plants Animals (vertebrate) Animals (vertebrate and invertebrate) Animals (invertebrate) Algae 80 A total of 53 viral families were observed for the Yuma lettuce virome. The viral family contig distribution for iceberg and romaine lettuce can be viewed in a phylogenetic tree constructe d using MEGAN software (Figure 18 ). The number of assigned contigs in this figure is represented by circle size and colors show the proportion belonging to each sample type. A large proportion of viruses were assigned to dsDNA Myovir idae, Podoviridae, and Siphoviridae bacteriophage fa milies of the C a u dovirales order as well as DNA Cau limoviridae and RNA Retroviridae retro - transcribing virus families which infect plant and vertebrate animal hosts respectively. The distribution of viral contigs associated with famili the 7 Yuma lettuce viromes was further analyzed (Table 20 and Figure 19). Of the v iral families % of the Yuma lettuce virome, C a u limoviridae and Retroviridae fa milies composed on average 25.6% and 14.7 % the iceb erg virome respectively. Similar results were observed for the romaine lettuce virome, with a large proportion of sequences assigned to Cau limoviridae (21.8 %) and Retroviridae (14.3 %) viral families . Myoviridae, Podoviridae, and Siphoviridae bacteriophag e families of the Cau dovirales order represented 11.6 - 14.6 %, 5.4 - 12.5 %, and 11.3 - 24.3 % of the iceberg lettuce (control, worker harvest, and post worker break) virome , respectively. Although the Caudovirales order represented 59.5 % of the romaine c ontrol le ttuce virome, only 2.5 - 8.7%, 2.3 - 7.5 %, and 3.2 - 11.1 % of viral contigs were assigned to Myoviridae , Podoviridae , and Siphoviridae viral families respectively for the remaining romaine lettuce samples (worker harvest, chop and wash, mixed salad). Other well represented viral families include other bacteriophage families ( Microviridae ) and viruses that infect plants ( Closteroviridae and Endornaviridae ), fungi ( Partitiviridae and Totiviridae ), algae ( P hycodnaviridae ), and amoeba ( Mimiviridae ). 81 Fi gure 18. Yuma lettuce viral family distribution 82 Table 20. Lettuce viral family distribution representing greater than 3.0% of the virome Host Sample Type * IC # contigs IC (%) IWH # contigs IWH (%) IPWB # contigs IPWB (%) Ice Avg (%) RC # contigs RC (%) RWH # contigs RWH (%) RCW # contigs RCW (%) RMS # contigs RMS (%) Rom Avg (%) Caulimoviridae 238 17.0 357 30.1 200 29.7 25.6 189 11.1 168 20.6 157 33.1 158 22.4 21.8 Closteroviridae 87 6.2 69 5.8 28 4.2 5.4 120 7.0 65 8.0 31 6.5 53 7.5 7.3 Endornaviridae 9 0.6 45 3.8 25 3.7 2.7 74 4.3 63 7.7 19 4.0 52 7.4 5.9 Microviridae 56 4.0 15 1.3 4 0.6 2.0 12 0.7 13 1.6 2 0.4 5 0.7 0.9 Mimiviridae 68 4.9 53 4.5 36 5.3 4.9 38 2.2 25 3.1 12 2.5 31 4.4 3.1 Myoviridae 205 14.6 150 12.6 78 11.6 13.0 400 23.5 71 8.7 12 2.5 56 7.9 10.7 Pa rtitiviridae 18 1.3 16 1.3 7 1.0 1.2 79 4.6 48 5.9 25 5.3 26 3.7 4.9 Phycodnaviridae 74 5.3 40 3.4 27 4.0 4.2 29 1.7 26 3.2 13 2.7 31 4.4 3.0 Podoviridae 175 12.5 64 5.4 42 6.2 8.0 188 11.0 51 6.3 11 2.3 53 7.5 6.8 Retroviridae 103 7.4 193 16.3 137 20.4 14.7 61 3.6 99 12.2 130 27.4 98 13.9 14.3 Siphoviridae 340 24.3 134 11.3 80 11.9 15.8 427 25.0 90 11.1 15 3.2 78 11.1 12.6 Totiviridae 28 2.0 51 4.3 9 1.3 2.5 88 5.2 95 11.7 48 10.1 64 9.1 9.0 * Iceberg control (IC), iceberg worker harvested (IWH), ice berg post worker break (IPWB), iceberg average (ice avg), romaine control (RC), romaine worker harvested (RWH), romaine chop and wash (RCW), romaine mixed salad (RMS) , romaine average (rom avg). 83 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Iceberg- Control Iceberg-Worker Harvest Iceberg- Post Worker Break Romaine- Control Romaine- Worker Harvest Romaine- Chop and Wash Romaine- Mixed Salad Contig Distribtuion Sample Figure 19. Lettuce viral family distribution representing greater than 3.0% of the virome Totiviridae Siphoviridae Retroviridae Podoviridae Phycodnaviridae Partitiviridae Myoviridae Mimiviridae Microviridae Endornaviridae Closteroviridae Caulimoviridae N=8 N=8 N=5 N=8 N=5 N=3 N=5 84 Yuma lettuce samples were further investigated at the viral species level to identify human enteric viruses. Rotavirus A and rotavirus C were identified in YL10 (iceberg worker harvest) and YL4 (romaine lettuce control) samples , respectively. In addition, picobirnavirus was identified in the iceberg control (YL18), romaine worker harvest (YL33 and YL42), romaine chop and wash (YL35), and romaine mixed salad (YL37) lettuce samples. The contigs belonging to these enteric viruses were further blasted against the NCBI nucleotide database (BL ASTn) to determine the gene, contig length, % identity, Query coverage (% of sequence that overlaps the subject sequence), E - value (number of hits expected by chance when searching a particular database) and viral host (Table 21). The rotavirus A contig s hared a 99% sequence identity similarity to the VP1 gene of a human isolate. Rotavirus C identified in the romaine lettuce control sample shared a 99% sequence identity similarity to the NSP2 gene of a bovine isolate. A majority of the picobirnavirus co ntigs had no significant similarity to any sequences in the BLASTn database. Two contigs belonging to the YL33 sample shared a 80% and 95% sequence identity similarity to the RDRP gene of human and porcine isolates respectively, however, the query coverag e for both picobirnavirus contigs was low (28 - 33%). Table 21. Yuma lettuce enteric virus nucleotide BLAST results Virus Host Sample Contig Number Gene C ontig length (bp) % Identity % Query coverage E - value Score Rotavirus C Bovine YL4 824 NSP2 483 99 79 0 686 Rotavirus A Human YL10 539 VP1 476 99 99 0 837 Picobirnavirus Human YL33 1487 RDRP 315 80 28 6.00E - 21 111 Picobirnavirus Porcine YL33 1133 RDRP 436 95 33 5.00E - 58 235 85 5. Discussion Before the development of NGS technology, virus detectio n in food matrices was primarily performed using one of two methods: cell culture or reverse transcriptase PCR (RT - PCR). Improvements in NGS technology have led to the ability to view entire microbial communities in a wide range of environments, providing the opportunity to investigate the role of viruses in various ecosystems including foods. However despite our current knowledge of fresh produce as a vehicle for human pathogenic virus transmission, there are no studies that have examined the virus commu nities in leafy greens or irrigation water which is a suggested pre - harvest source of fresh produce contamination. In this study we developed a metagenomic pipeline to investigate the virome from iceberg and romaine le ttuce from the field as well as from irrigation water used to cultivate these crops. Virus elution and concentration prior to extraction and sequencing are critical steps in the metagenomic pipeline that ultimately impact virus recovery. In our study we evaluated a method of virus recover y from lettuce using an elution protocol adapted from Dubois et al., ( 2006 ) and enumeration by double agar overlay plating. Viruses such as MS2 are thought to use physicochemical forces, specifically electrostatic forces, for nonspecific attachment to soli d surfaces such as fresh produce and a higher pH has been shown to effectively dissociate viruses from lettuce surfaces (Deboosere et al., 2012, Vega et al., 2008, Vega et al. 2005) . P22 re covery from lettuce ranged 43.3% - 77.4 %, while MS2 had a broader ran ge of re covery between experiments (2.8% - 91.1 %). The low recovery was due to trial 2 and could have been due to issues during enumeration using double agar overlay plaque assay, which is subjective to diluting, plaque size, incomplete lysis, or plaque aggr egation leading to inaccurate measurements ultimately variation in virus recovery (as described in Kropinski et al., 2009) . Other possibilities 86 include incorrect dilution spiking suspension or a potential poorly prepared virus stock. However, MS2 recover y in trials 1 and 3 in this study (29.1 - 91.1 % ) was similar to a study by Dubois et al., ( 2006 ) who achieved a 22.9 - 96.2% recovery of MS2 from inoculated butter lettuce using the same elution protocol. For comparison, s tudies investigating enteric (norovir us, hepatitis A, poliovirus) and enteric surrogate (murine norovirus) virus recovery from fresh produce using a real - time RT - PCR method have discovered virus recovery can range anywhere between 8.2% and 100%, and can vary within and between viral species (Hennechart et al. , 2002; Sánchez et al., 2012) . A total of six 100 - L irrigation water samples as well as 21 iceberg and 21 romaine lettuce samples were collected for metagenomic analysis. Yuma irrigation water had low turbidity and low levels of total coliforms as well as enterococci and E. coli which are the typical fecal indicator bacteria used in water quality analysis. Viral communities were sequenced using Illumina Hiseq technology and were analyzed using a variety of b ioinformatics tools targeted to short sequencing reads. For both the Yuma irrigation water and lettuce viromes, a large proportion of viruses had no hits against the viral NCBI viral database. This observation is consistent with most other viral metagenomic studies to date (Angl y e t al., 20 06, Aw et al., 2014 , Breitbart et al., 2002, Mokili et al., 2012, Rosario et al., 2009) . No hits could represent either novel viruses or generally poor global data bases for viral genome comparison. These data suggest , however that we have limit ed knowledge of the viral genome diversity in these environments and that there is a strong need to develop and improve current metagenomic techniques and viral databases. Of the irrigation water contigs that were assigned, a total of 64 viral families were identified with dsDNA viruses dominating the virome. This is largely due to the high recovery 87 of Myoviridae , Podoviridae , and Siphoviridae bacteriophage virus families of the Cau dovirales order. Viral metagenomic studies of sewage, reclaimed water, human feces, and fermented foods have shown that these viromes also contain a large proportion of bacteriophage viruses belonging to the Cau dovirales order using either 454 pyrosequencing or Illumina technology (Aw et al., 2014 , Kim et al., 2011, Park et a l., 2011, Rosario et al., 2009) . Bacteriophage that infect E. coli are promising indicators of fecal pollution and have even been suggested as possible bacterial control agents in foods, however little is known about their role within virus communities in irrigation water and fresh produce systems (Dawson et al., 2005; Havelaar et al., 1993; Sharma, 2013) . Investigation of the host distribution of these phage species showed that most of the bacterial host species were common to freshwater or marine enviro nments ( Pelagibacter , Cellulophaga, Synechococcus, Bdellovibrio , Bacillus , Pu niceispirillium , Mycobacterium , and Prochlorococcus ) with low relative abundance of phage infecting the genus Escherichia or Enterococcus which are typical bacterial fecal indicat ors. This is similar to our enterolert and colilert results and may indicate low levels of fecal pollution in irrigation water; however , these bacterial indicators fail to recognize human enteric viral hazards which have been shown not to correlate with t ypical bacterial indicators in water (Harwood et al., 2005) . This is one of the first studies to u se metagenomics to investigate viral species in irrigation water and lettuce and there is interest in how these data may be of i mportance to the food indust ry. Analysis of viral species in irrigation water revealed numerous viruses infecting agricultural insect pests, crops, fish and shrimp, livestock animals, and even humans. Although crop viruses identified were not specific to lettuce, many plant viral s pecies belonging to Geminiviridae, Nanoviridae, and Tombusviridae families can cause disease in multiple food crops and are of interest to the fresh produce industry due to h igh economic costs from reduced 88 quality, productivity, or yield (Boulila 2011, Var ma and Malathi 2003) . Animal virus families were well represented by the Circoviridae and Poxviridae families. These plant and animal virus families have also been identified in wastewater and reclaimed water using next - generation sequencing (Aw et al., 2014, Rosario et al., 2009) . The identification of human enteric viruses is of particular importance to public health and monitoring food safety. In this study picobirnavirus , coronavirus , parechovirus , and rhinovirus as well as hepatitis A and E were human pathogens identified in irrigation water. Hepatitis A is an enteric viral pathogen that has been linked to fresh produce outbreaks while hepatitis E is just recently gaining recognition as a food safety hazard. Hepatitis E is transmitted through th e fecal - oral route and genotypes 3 and 4, which are zoonotic, are now suspected of foodborne transmission (e.g. pig meats) (Meng, 2011) . The picobirnavirus identified is also of interest because this dsRNA virus is likely zoonotic, is a suggested cause of diarrhea in pigs, and has been also isolated from the feces of immunocompromised and healthy individuals (Ganesh et al., 2012) . However, the low percentage of sequence identity similarities for picobirnavirus as well as a majority of the plant, animal, a nd insect viral species detected sugg ests these are novel viruses based on our limited knowledge of the viral diversity in irrigation water. For the iceberg and romaine lettuce viomes, dsDNA viruses also composed a large percentage of assigned contigs with a majority belonging to bacteriophage families of the Cuadovirales order as well as Phycodnaviridae and Mimiviridae families which infect a lgae and amoeba , respectively. Interestingly, iceberg and romaine lettuce con tained of a greater proportion of retro - transcribing families, including C a u limoviridae (DNA) and Retroviridae (RNA). Retro - transcribing viruses use a reverse transcription step in order to replicate forming a DNA from an RNA strand . For viruses in the Retroviridae family , infection occu rs following 89 incorporation of viral DNA into the host s cell genome however viruses of the Caulimoviridae family do not require integration and replication is instead in the cytoplasm. The Caulimoviridae family is a group of viruses that is widely distribu ted but known to cause serious crop and plant disease in tropic regions (Geering, 2007) . Retroviridae species are known to infect humans and animals; however viruses within this family are not recognized as foodborne pathogens. Interestingly , control iceb erg and romaine lettuce samples collected by researchers had a larger proportion of phage viruses when compared to subsequent stages of field harvest. It is unclear as to why control samples had more viruses infecting bacteria. Assuming phage correspond with bacterial presence, bacteria could have been removed from romaine lettuce during chop and wash resulting in lower populations in these samples. However, this does not explain the lower levels observed following worker harvest or in iceberg lettuce (c ontrol vs. worker harvest) which is not washed prior to packaging. Metagenomics is a random sequencing approach and combined with a small sample size, variations in virus communities between samples are likely. Further studies with a larger sample size as well as use of a quantification assay (qPCR) could be helpful to investigate differences between these sample types. Interestingly, dsRNA viruses including Rotavirus A and human picobirnavirus were discovered in the lettuce virome. Rotavirus A, whic h can be transmitted through food including fresh produce, is a common cause of gastroenteritis in children but has also been shown to cause disease in adults (Fletcher et al., 2000, Newell et al., 2010) . Rotavirus A was found in an iceberg lettuce sample collected by workers and found to have a high percentage identity match to a human isolate, which suggests this enteric virus was present on the lettuce either prio r or following field harvest. In addition, r otavirus A is a well - known waterborne enteric pathogen and possible sources of co ntamination could be from fecally 90 contaminated irrigation water or poor worker hygiene. Bovine rotavirus C was identified in a r omaine lettuce control sample, indicating fecal co ntamination from cattle with possible sour ce s including contaminated irrigation water or runoff or improperly treated manure used as fertilizer. However the location of cattle farms near the farm as well as whether manure was used to fertilize the fields is unknown. H uman picobirnavirus was also identified in both iceberg and romaine lettuce. These data suggest that dsRNA enteric viruses pose a threat to food safety beginning at the pre - harvest stage of production. However, a majority of picobirnavirus contigs were n ot significantly similar to sequences in the NCBI nucleotide database suggesting limited knowledge of this virus genome and the role it pl ays in water and food systems . In addition, metagenomics detects DNA or RNA and it is unknown as to whether these organisms are viable and infect ious. M etagenomic analysis revealed that most of the viruses in both irrigation water and lettuce are novel or unknown. Of those assigned, a total of 47 viral families were shared between the irrigation water and lettuce viromes. Analysis of v iral fami 0% of the irrigation water and lettuce virome s revealed six viral families (Microviridae, Mimiviridae, Myoviridae, Phycodnaviridae, Podoviridae, Siphoviridae ) in common. As stated previously, viral families of the Caudovirales order a re numerous in a variety of environmental samples. Interestingly, the Phycodnaviridae viral family infec ts algae or aquatic chlorophyll containing organisms. This indicates the presence of algal viruses in the lettuce virome, which may be a result of the irrigation water used during cultivation. However, the role of these viruses in lettuce and the lettuce virome is currently unknown. Viruses infecting animals (vertebrate and/or invertebrate) and plants in irrigation water and lettuce varied in proportio n and viral family type. Yuma lettuce contained a larger proportion of plant viruses (19.16 - 39.15%) compared to 91 irrigation water (0.95 - 7.17%) , largely due to the Endornaviridae , Closteroviridae, and Caulimoviridae viral families. In addition, of the anim al 0% of the irrigation water virome , a large proportion belonged to those infecting both vertebrates and invertebrates ( Poxviridae and Circoviridae families) while in the lettuce virome a majority of viruses belonged to the Ret roviridae family which infects vertebrates alone. Of the human viruses identified, dsRNA picobirnavirus was the only enteric virus identified in both lettuce and irrigation water . This virus, which had low percentage identities, needs to be further inves tigate d for further potential involvement in water and food systems. These data show that the same families of algal viruses , phage, and human enteric viruses can be present in both the lettuce and irrigation water virome s , while animal viruses tend to di ffer between the se two environments. 92 I V . VIRAL CROSS - CONTAMINATION OF LET TUCE D URING SMALL - SCALE LEAFY GREEN P ROCESSING 1. Introduction Fresh produce consumption in the United States has increased in recent years as consumer s desire to main tain a healthy lifestyle . Consequently, foodborne outbreaks associated with fresh produce are becoming increasingly recognized. Leafy greens in particular are often consumed raw and considered a food commodity of high risk, accounting for the highest num ber of foodborne illnesses (2.2 million) between 1998 and 2008 (Painter et al., 2013). Furthermore foodborne outbreak data from this same time period suggests viruses are commonly associated with leafy green contamination and transmission, with norovirus and leafy greens the pathogen - commodity pair most likely to be associated with a foodborne outbreak (Gould et al. 2013). Although human norovirus has been frequently associated with leafy green outbreaks, there is currently limited knowledge on viral con tamination in the supply chain and other viral pathogens of concern. In response to the changing fresh produce supply and demand, traditional agricultural and post - harvest practices have been altered which ultimately increase food supply chain complexity. Practices such as cutting and coring at harvest, increased importation and transportation, and large scale production facilities are now employed to support changing consumer habits (Heaton et al. 2008, Lynch et al. 2009). These practices contribute to the increasing number of leafy green outbreaks by providing multiple opportunities for human viral pathogen contamination and ultimately make it difficult to determine the single source of contamination. At the post - harvest level of leafy green production , contamination can occur during any stage of processing, packing, storage, or transportation. Critical control points include the quality of water used 93 (cooling, washing), worker hygiene, and the condition or overall cleanliness of processing equipment, cooling facilities, storage and packaging containers, and transportation vehicles ( CFSAN , 2006 ) . Leafy greens are now provided to the consumer either as bulk products to be washed (e.g., head of lettuce) or as ready - to - eat (RTE) salads. Many consumers today desire RTE salads which are typically shredded, washed, dewatered, dried, and packaged prior to cold storage and transportation to distribution centers. Leafy green washing is an important step for improving the quality of the food product by removi ng contaminants (soil, debris, microorganisms) and prolonging shelf life. In addition, proper disinfection and monitoring of sanitizer levels during processing of RTE salads are essential steps to minimize foodborne disease (CFSAN, 2014 ) . Although numero us physical and chemical disinfection methods have been investigated for use in foods, washing with a chlorine - based sanitizer is the disinfection method most commonly used in the fresh - cut produce industry due to the low cost and limited negative impact o n product quality (CFSAN , 2014) . Guidelines for washing fresh produce in chlorinated water include a maximum free chlorine (hypochlorite) concentration of 200 ppm and generally a 1 to 2 minute contact time (CFSAN , 2014) . Currently a 5 - log pathogen reduct ion standard is suggested for fresh produce production, however, this standard primarily targets bacteria l pathogens and to date there is no defined criterion for antiviral disinfectants (Allwood, Malik, Hedberg, & Goyal, 2004; Gulati, Allwood, Hedberg, & Goyal, 2001 ) . There have been numerous studies investigating bacterial pathogen reduction on fresh produce using chlorine sanitizers. Pilot - scale studies have found that chlorine - based sanitizers generally reduce bacterial pathogen populations on lettuc e between 1 and 3 logs (Davidson et al., 2013; Gil et al. , 2009) . Several laboratory studies investigating the effects of chlorine on viruses 94 when inoculated onto fresh produce have also shown that viral (MS2, feline calicivirus, murine norovirus) reducti on generally does not exceed 3 logs when exposed to a variety of free chlorine (15 - 800 ppm) levels (Allwood et al., 2004, Dawson et al., 2005, Fraisse et al., 2011, Gulati et al., 2001 ) . However, it is difficult to deduce how these data relate to large - sc ale leafy green processing and currently there is limited research investigating the bacterial and viral reduction on fresh produce during simulated commercial processing. 2. Research Goals and Objectives The ultimate goal of th e second half of the thes is research was to i nvestigate the efficacy of current post - harvest leafy green processing and disinfection practices to better understand viral risks from farm - to - fork during a contamination event . Specifically, the goal was to a ssess the efficacy of a c hlorine - based sanitizer against coliphage MS2 ( an enteric virus surrogate ) on romaine lettuce during simulated commercial processing . Having found a wide variety of viruses on lettuce from the field , there was interest in how removal might be achieved du ri ng processing after harvest . The stu dy objectives were as followed: 1. Evaluate MS2 reduction on romaine lettuce during and following small - scale commercial leafy green processing (shredding, flume washing, shaker table dewatering, centrifuge drying) with an d without a sanitizer wash treatment 2. Determine MS2 levels in the flume wash water and centrifugation water following processing 95 3. Materials and Methods 3.1 Bacteriophage Inactivation and Free Chlorine Demand 3.1.1 MS2 Inactivation during Sanitizer Exp osure Prior to small - scale commercial leafy green processing trials, experiments were performed on the bench to determine the extent of MS2 ina ctivation when exposed to a chlorine - based sanitizer ( XY12 Ecolab, St. Paul, MN, USA) containing 25 ppm of avai lable chlorine as recommended on the manufa fruit and vegetable washing. Previous viral inactivation studies have used similar co ncentrations of free chlorine (Casteel et al., 2008; Casteel, Schmidt, & Sobsey, 2009; Fraisse et al., 2011) . To achieve a level of 2 5 ppm available chlorine in the solution, 31.3 µL of XY12 were added to 100 mL of distilled water in a 250 mL glass bottle and the pH was adjusted to 7 using 6 N HCl . The sanitizer solution was mixed and the available chlorine con centration was confirmed using a chlorine color disc test kit (Hach test kit, 0.0 - 3.5 mg/l Model CN - instructions. Since the available chlorine concentration was out of the test kit measurement range, the solution was diluted 10 - fold in PBW before analysis. Following free chlorine confirmation, 1 mL of high titer MS2 (10 10 PFU/mL) was added to the sanitizer solution and vortexed. After 30 s, 60 s, and 120 s of exposure, 1 mL of sample was transferred t o a 500 mL plastic bottle (one for each time point) containing 250 mL of tris - glycine buffer (pH 9.5) and 2% sodium thiosulfate (mimicking the virus elution recovery assay). The pH was adjusted to 7.2 ± 0.2 and the samples were diluted in PBW before platin g using the double agar overlay met hod described in Part III, S ection 3.1.2, pages 37 - 38. The free chlorine level in the neutralized samples was then measured using the same color disc test kit described above. As a control, 1 mL of MS2 (10 10 PFU/mL) was added to 100 mL of distilled water without sanitizer. The 96 control was processed following the same protocol as the sanitizer challenge experiment. In addition, the MS2 suspension used for inoculation was diluted in PBW and plated to determine the inoculu m level . The plaque concentration for the inoculated suspension, control and challenge samples was calculated using Formula 1 from Part I II , Section 3.1.3, page 39. 3.1.2 Coliphage MS2 Chlorine Demand To test the chlorine demand of the bacteriophage duri ng chlorine inactivation, MS2 was exposed to a stock solution containing 25 - ppm available chlorine . One mL of stock, TSB - based MS2 (10 10 or 10 7 PFU/mL) or PBW diluted MS2 (10 8 or 10 6 PFU/mL) was added to a 250 - mL glass bottle containing 25 ppm of availabl e chlorine (achieved by adding 31.3 µL of XY12 to 100 mL of distilled water). The available chlorine was measured before the addition of phage and after 0.5 min, 1 min, 3 min, 5 min, 10 min, and 15 min of exposure using the color disc test kit described in the bacteriophage inactivation experiment above. If necessary, the solution was diluted 10x in PBW before free chlorine measurement and analysis. 3.2 Bacteriophage Reduction during Small - Scale Leafy Green Processing With and Without a Chlorine - Based San itizer 3.2.1 Lettuce and Processing Line P reparation Heads of romaine lettuce were purchased from a local produce supplier the day before processing. Upon arrival, the core of each romaine lettuce head was removed by cutting 2.5 to 5 cm from the core us ing a sterile scalpel. Remaining romaine l ettuce leaflets were weighed until a total of 6 - kg was achieved. The 6 - kg batch was placed into Whirl - pak bags and stored at 4°C unti l processing the following day. A small - scale commercial leafy green processin g line located in the Department of F ood Science and Human Nutrition at Michigan State University was used (under the supervision of Dr. Elliot Ryser). The processing line included a lettuce shredder 97 ( TranSlicer 2500, Urschel, Valparaiso, IN ) , conveyor ( m odel 736018 mc series, Dorner Manu - facturing, Hartland, WI ) , flume tank ( 3.6 m; Heinzen Manufacturing, Inc., Gilroy, CA ) , mechanical shaker table ( Baldor Electric Co., Ft. Smith, AR ) , and centrifugal dryer ( model SD50 - LT, Heinzen Manufacturing, Inc. ) as d escribed by Buchholz et al. (2012). For lettuce washing, water was recirculated through a stainless s teel water recirculation tank (1,000 L volume ) connected by a hard plastic discharge hose to the stainless steel flume tank containing two overhead spray jets using a centrifugal pump (model XB754FHA, Sterling Electric, Inc., Irvine, CA) at 15 liters/s (Buchholz et al. 2012) . The proc essing line sampling locations are shown in Appendix , Figure A4 . The day before use, the entire processing line as well as t he 121 - L collection and inoculation bins were sanitized by spraying with 200 ppm of Quorum Clear (sodium hypochlorite active ingredient) ( Ecolab, St. Paul, MN, USA ). After at least a 1 min exposure all equipment was rinsed with tap w ater and air - dried over night. 3.2.2 L ettuce Inoculation and Sampling Prior to running the processing experiment, un - inoculated lettuce was sent through the shredder in order to prime the machine. Three ~50 - g samples of the un - inoculated shredded lettuce were taken as negativ e control s . Prior to lettuce inoculation the water recirculation tank was filled with 800 L of municipal tap water ( Michig an State University <0.05 ppm of free chlorine). For sanitizer ex periments, approximately 946.4 mL of XY12 was added to the 8 0 0 L of water to achieve a free chlorine concentration of 25 ppm. The free chlorine level in the flume wash water was confirmed using the same chlorine color disc test kit mentioned previously and by adding small volumes of hydrochloric acid to the flume wash water as it was recirculated through the flume and holding tank. For inoculation, a 6 - kg batch of lettuce w as submerged for 15 min in a sanitized 121 - L plastic bin containing 80 L 98 of municipal tap water ( Michig an State University <0.05 ppm of free chlorine) and 100 m L of MS2 previously grown to 10 9 - 10 10 PFU/mL. Th e lettuce was then placed into the dewatering centrifuge to remove excess water prior to processing and three ~50 - g samples were collected. The remaining virus inoculated lettuce was then sent through the e nti re processing line with three ~50 - g samples taken at four different stages of processing. Samples were collected after shredding , flume washing (2 min), shaker table dewatering, and centrifugal drying following processing. At each station, designated workers wearing latex gloves placed a handful of lettuce into a Wh irl - pak bag then weighed out 50 - g samples which were transferred to a fi lter Whirl - pak bag. In addition, samples of centrifugation water and flume wash water were collected post - processing. Approximately 500 mL of water was collected fr om the centrifuge drain during the last stage of lettuce drying for both sanitizer - free and sanitizer experiments. For sanitize r experiments alone a total of two 20 L wash water samples were collected from the flume tank and further concentrated using hollow - fiber ultrafiltration. The flume tank wash water samples were collected to confirm efficacy of the sanitizer. 3.2.3 Lettuce Processing and Plating Lettuce was processed to recover the phage using the el ution proced ure described in Part III, Section 3.1.2, pages 37 - 38 . Each 5 0 - g lettuce sample was eluted with 250 mL of tris - glycine buffer at pH 9.5 and shaken for 20 min with the e luent pH adjusted to 7.2 ± 0.2. For the sanitizer experiment , lettuce was immediately transferred to Whirl - pak bags containing 250 mL of buffer and 2% sodium thiosulfate following flume wash ing, shaking , and final centrifugal drying sampling. In addition, 2% sodium thiosulfate was added to flume wash water and centrifugal water sampled during sanitizer processing. The lettuce eluent and water samples were plated using the double agar overlay method described in Part III, Sections 3.1.1 and 3.1.2, pages 35 - 38. 99 3.2.4 Calculations and Statistical A nalysis The plaque concentrat ions for the inoculation suspension and recovered eluent from each ~50 - g lettuce sample collected immediately after inoculation, after shredding, flume washing, shaking, and centrif uging were calculated using formula 1 ( Par t III, Section 3.1.3, page 39 ). In addition, the c oncentration of phage per g of romaine lettuce (PFU/g) was calculated using formula 2 ( Part III, Section 3.1.3, page 39 ). To estimate the plaque concentration of the 80 L inoculation suspension, the inoculated concentrate (PFU/mL) was m ultiplied by the volume of inoculant concentrate added (100mL) which was then divided by the total volume of water (80000 mL) as shown in formula 5. The estimated concentration per g of lettuce following inoculation (but prior to centrifugal drying) was c alculated by multiplying the estimated plaque concentration in the inoculation suspension by the total volume of the suspension (total phage in inoculation suspension) and dividing by the total weight of lettuce added to the suspension (6000g) (formula 6). An unpaired two sample t - test was used to compare virus removal during th e various stages of processing ( P 0.05). For spent centrifuge water and flume water samples, the MS2 concentration (PFU/mL) was ca lculated using formula 1 ( Part III, Section 3.1.3, page 39 ) . The total viral reduction (PFU) in the flume water before and after washing was estimated by s ubtracting the total MS2 entering the flume wash water by the total MS2 exiting the flume wash water. Total MS2 ente ring the flume wash water was calculated by first calculating the total PFU on lettuce entering the flume (multiply the average PFU/g after shredding by the total lettuce ( 5700 g)) and subtracting the total PFU on lettuce exiting the flume (multiply the average PFU/g after flume washing by total lettuce ( 5700 g)). Total MS2 remaining in the flume wash water was calculated by first estimatin g the total PFU/mL in the 20 L sample (PFU/mL multiplied by eluent concentrate volume). This value was then divided by our sample volume 100 (20 L) and multiplied by the total wash water volume (800 L) to get the remaining PFU in the flume wash water (formula 7) . (5) 4. Results 4.1 MS2 Inactivation during Sanitizer Exposure A benchtop inactivation experiment wa s conducted prior to small - scale leafy green processing trials to determine the relative inactivation of MS2 upon exposure to 25 ppm of free chlorine . One mL of high titer (10 10 PFU/mL) TSB - based MS2 was added to the sanitizer solution (100 mL) and sample s were processed after 30 s, 1 min, and 2 minutes of exposure. The time series was chosen based on standard commercial l ettuce wash in g and exposure practices . For the liquid suspension challenge experiment, all samples started with 25 ppm of free chlorin e and were immediately neutralized (0 ppm) with 2% sodium thiosulfate after the specific contact times. Figure 20 shows the log transformed total phage values at time 0 (represented by the control) and after 30 s, 1 min, and 2 min of exposure to 25 ppm fr ee chlorine. Compared to the initial population , the number of phage decreased 2.2, 1.2 , and 2.4 log s after 0.5, 1, and 2 min time points , respectively , with an o verall average reduction of 1.9 log s . 101 4.2 Coliphage MS2 Chlorine Demand The phage prepara tion was not purified and thus may carry a chlorine demand. This was evaluated and F igure 21 shows the free chlorine demand after adding 1 mL of MS2 phage stock (suspended in either TSB or PBW) to 100 mL of the wash solution (~2 fold dilution) . After 15 min, the free chlorine level remained unchanged or only dropped 1 ppm when exposed to 1 ml of high titer (10 6 and 10 8 ) PBW - based MS2 in 100 mL of sanitizer. After 30 s of exposure to 1 mL TSB - based MS2 at 10 10 and 10 7 (PFU/mL), free chlo rine levels were re duced to 0.3 and 0.5 ppm respectively. After 15 min , free chlorine leve ls were reduced to 0.2 and 0.1 ppm , using 10 10 and 10 7 (PFU/mL) TSB - based MS2 respectively. These data show that TSB exhibited a demand on for free chlorine. This could be similar to an actual virus contamination event in which the virus is associated with fecal matter . 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 0 0.5 1 1.5 2 2.5 Log (total MS2) Time (min) Figure 20. Coliphage MS2 inactivation after exposure to 25 ppm free chlorine in liquid suspension 102 * Green line representing 10 10 TSB trial is located behind purple line representing 10 7 TSB trial 4.3 Virus Reduction during Leafy Green P rocessing Small - scale e xperiments were conducted with and without sanitizer to determine virus reduction on lettuce during simulated commercial processing. Lettuce was inoculated by immersion in 80 L of municipal tap water containing 100 mL of TSB - based MS2 (10 9 - 10 10 PFU/mL) an d examined prior to processing (after inoculation and drying) as well as following shredding, 2 min of flume washing, sh aker table dewatering , and centrifugal drying. TSB - based MS2 (100 mL) was diluted ~3 fold when added to tap water (80 L) and was theref ore not diluted in PBW prior to inoculation. For lettuce washing, a chlorine based sanitizer was used according a negative control. The first sanitizer trial was the only trial (including trials with and without sanitizer) to have a positive result in the negative control lettuce. A positive result was observed 0.0 5.0 10.0 15.0 20.0 25.0 30.0 0 0.5 1 3 5 10 15 Free Chlorine (ppm) Time (min) Figure 21. MS2 free chlorine demand 10^8 (PBW) 10^6 (PBW) 10^10 (TSB) 10^7 (TSB) 103 for all three lettuce samples (triplicates) when plated undilute d on TSA overlay plates, with no mo re than 5 plaques on a plate and an average concentration of 3.5 PFU/g. This indicates either cross contamination d uring sampling or plating or the presence of E. coli Famp on the lettuce prior to processing. The MS2 concentrations in the original cult ure, the inoculation suspension , and estimated MS2 concentration on the lettuce prior to drying are provided for both sanitizer and non - s anitizer experiments in Table 22 . The MS2 concentration in the inoculation suspension was approximately 10 7 and 10 6 PF U/mL with average MS2 levels on romaine lettuce of 8.5 log and 7.8 log (PFU/g) for sanitizer free and sanitizer trials , respectively. The average MS2 concentrations on romaine lettuce (PFU/g) for triplicate lettuce samples collected before processing and after s hredding, flume washing, shaker table dewatering , and centrifugal drying for both sanitizer and non - sanitizer exp e riments are presented in Table 23 . For all three trials, the average MS2 concentration on romaine lettuce before processing (following inoculation and drying) was 6.0 and 4.9 log (PFU/g) for the sanitizer - free and sanitizer experiments , respectively. Following shredding, flume washing, sh aker table dewatering , and centrifugal drying, the average MS2 concentration decreased to 5.0 and 4. 2 log (PFU/g) for sanitizer - free and sanitizer experiments , respectively. Free chlorine levels tested following processing showed a reduction of 10 ppm, 5 ppm, and 5 ppm for trials 1, 2, and 3 respectively. 104 Table 22 . Estimated MS2 concentration in inoculation suspension and on r omaine lettuce following inoculation (prior to centrifugal drying) Wash w ater Trial Original culture concentrate (Log PFU/mL) Estimated MS2 concentration in inoculation suspension (Log PFU/mL) Estimated MS2 concentration on lettuce ( Log PFU/g) Average estimated MS2 concentration on lettuce ( Log PFU/g) Sanitizer - f ree 1 10.4 7.5 8.7 8.5 2 10.4 7.4 8.6 3 10.1 7.2 8.3 Sanitizer 1 9.5 6.6 7.7 7.8 2 9.5 6.6 7.7 3 9.7 6.8 7.9 Table 23 . Average MS2 concentration on romaine lettuce for triplicate lettuce samples collected following various stages of small - scale leafy greens processing Flume wash water Trial Average Concentration of MS2 on r omaine lettuce (log PFU/g) Before p rocessing Shredder Flume w ash Shaker Cen trifuge Sanitizer - free 1 6.1 5.8 5.2 4.9 5.1 2 5.9 5.8 4.9 4.7 5.0 3 5.9 5.7 4.8 4.9 4.9 Average 6.0 5.8 5.0 4.8 5.0 Sanitizer 1 5.0 4.9 4.4 4.4 4.5 2 4.9 5.0 3.9 3.9 4.0 3 4.8 4.9 4.2 4.1 4.0 Average 4.9 4.9 4.2 4.2 4.2 105 The MS2 concentra tions on romaine lettuce at various s tages of leafy green processing without and with the sanitizer treatment are shown in Figures 22 and 2 3 , respectively. For sanitizer - free experiments, the average lettuce phage concentrations before processing were 6 .1 , 5.9, and 5.9 log PFU/g for trials 1, 2 and 3 respectively. Following shredding, 2 min of flume washing ( tap water alone ), shaker table dewatering , and centrifugal drying, MS2 phage populations decreased to 5.0 , 5.0, and 4.9 log PFU/g for trials 1, 2 and 3 respectively, with an o verall average reduction of 1.0 log PFU/g. For experiments with 25 ppm of free chlorine included in the flume water , the average lettuce phage concentrat ions before processing were 5.1, 4.9, and 4.8 log PFU/g for trials 1, 2 and 3 respectively. Following shredding, 2 min of flume w ashing (with sanitizer), shaker table dewatering , and centrifugal drying, MS2 phage populations decreased to 4.5, 4.0, and 4. 0 log PFU/g for trials 1, 2 and 3 respectively, with an overall average reduc tion of 0.8 log PFU/g. Although centrifugal drying was the final lettuce processing step, the largest average reduction occurred following shaking, reducing phage populations by a total of 0.8 and 1.1 log PFU/g for experiments with and without sanitizer , respectively. In addition, coliphage MS2 numbers remained relatively stable following flume washing with or without 25 ppm free chlorine. No statistical difference was observed for the total phage reduction (log PFU/g) between water and sanitizer treatmen ts (P>0.05). 106 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 Before Processing Shredder Flume Tank Shaker Dewatering Centrifuge MS2 Log (PFU/g) Sampling Location Figure 22. MS2 reduction on inoculated romaine lettuce during small - scale leafy green processing without sanitizer Trial 1 Trial 2 Trial 3 0.0 1.0 2.0 3.0 4.0 5.0 6.0 Before Processing Shredder Flume Tank Shaker Dewatering Centrifuge MS2 Log (PFU/g) Sampling Location Figure 23. MS2 reduction on inoculated romaine lettuce during small - scale leafy green processing with sanitizer Trial 1 Trial 2 Trial 3 107 For both sanitizer and sanitizer - free treatments, the reduction of phage populations on romaine lettuce between consecutive stages of food processing was compared (Table 24). The largest reduction occurred between s hredding and flume washing, ranging from 0.5 to 1.0 log PFU/g and 0.6 to 1.0 log PFU/g for sanitizer and sanitizer - free treatments, respectively. For sanitizer - free experiments, phage populations decreased an average of 0.2 log PFU/g between before proces sing and shredding and 0.1 PFU/g between fluming and shaking. This is comparable to experiments with sanitizer where little reduction of MS2 on lettuce occurred before or following flume washing. Interestingly, lettuce phage populations for both sanitize r and sanitizer - free experiments increased slightly between shaking and centrifugation in all trials but one (trial 3 sanitizer). Table 24 . Log reduction of MS2 on r omaine lettuce between consecutive processing stages Flume Wash Water Trial Log removal (PFU/g) Before processing - Shredding Shredding - Flume Flume - Shaker Shaker - Centrifuge Sanitizer - free 1 0.4 0.6 0.3 - 0.2 2 0.1 1.0 0.1 - 0.3 3 0.1 1.0 - 0.1 - 0.1 Average 0.2 0.8 0.1 - 0.2 Sanitizer 1 0.2 0.5 0.1 - 0.1 2 - 0.1 1.0 0.0 - 0.0 3 - 0.1 0.8 0.1 0.1 Average 0.0 0.8 0.1 0.0 The MS2 conce ntration (PFU/mL) in centrifugation water decreased during lettuce drying following processing , ranging from 5.0 log PFU/mL to 5.5 log PFU/mL (average 5.2 log PFU/mL) and 3.8 log PFU/mL to 4.2 log PFU/mL ( average 4.0 log PFU/mL) for the sanitizer - free and sanitizer experiments , respectively. Two, 20 L flume water samples were also collected after sanitizer processing of lettuce . The average MS2 concentration in 20 L of flume wash water was 3.2 PFU/mL and d id not exceed 4.6 PFU/mL f or all three sanitizer trials. MS2 reduction due 108 to flume washing with sanitizer was es timated and included in Table 25 . For all three trials, the average total MS2 log (PFU) r eduction in the flume was water was about 4 logs. Tabl e 25 . F lume water MS2 levels and reduction following processing Trial PFU log (PFU) Total MS2 on lettuce entering flume Total MS2 on lettuce exiting flume Total MS2 lost on lettuce during flume wash Total PFU in 20 L of flume water Total MS2 in flume wa ter before washing Total MS2 in flume water remaining MS2 Removal 1 4.1x10 8 1.5 x10 8 2. 6 x10 8 8.0 x10 2 8.4 4.5 3.9 2 5.1 x10 8 4.8 x10 7 4.6 x10 8 1.2 x10 3 8.7 4.7 4.0 3 5.0 x10 8 8.4 x10 7 2.9 x10 8 8.1 x10 2 8.6 4.5 4.1 Avg 4.7 x10 8 9.3 x10 7 3.4 x10 8 9.3 x1 0 2 8.6 4.6 4.0 5. Discussion Viral foodborne diseases are a growing concern for raw p roduce. With consumer demand increasing for fresh - cut lettuce , this study was focused on determining if the use of chlorine - based sanitizers might provi de some abil ity to decrease virus on incoming lettuce. MS2 coliphage was used as an enteric virus surrogate on romaine lettuce during simula ted commercial processing. This phage is no n - enveloped, composed of ssRNA , and shares similar resistance characteristics to e nteric viruses including norovirus and can thus be used to represent foodborne viruses that cannot be propagated in cell culture (Dawson et al., 2005) . A small - scale laboratory experiment was conducted to determine the efficacy of a chlorine - based sanit izer against MS2 in liquid suspension. Th is experiment demonstrated a range in virus inactivation from 1.2 - 2.4 logs (PFU) and was comparable to a study by Doultree et al. ( 1999 ) . They examined the effect of 100 - 500 ppm of free chlorine (Det - Sol 5000 san itizer) using feline calicivirus and reported 1.75 to 2.75 log inactivation s in liquid suspen sion following a 1 min exposure. This suggests that the inactivation efficacy of 25 ppm free chlorine against MS2 virus in liquid suspension as observed in this study is com parable to inactivation rates for 109 other enteric virus surrogates at higher free chlorine levels (some exceeding maximum levels allowed in the food industry). One of the issues is the demand for the halogen within the food processing environme nt whic h leads to the depletion of free chlorine . A small - scale laboratory experiment conducted to examine the chlorine demand of high titer, TSB - cultured , MS2 demonstrated this phenomenon, finding that TSB - based MS2 had a high demand for free chlorine im mediately following exposure to 25 ppm w hile MS2 diluted in PBW did not . Previous studies have shown that organic compounds, including culture media and microbial suspensions, have a high demand for chlorine and reduce free chlorine levels following expos ure (Kotula et al. , 1997, Shang & Blatchley, 2001) . These data suggest that the presence of organic substrates in liquid suspension s strongly impacts sanitizer efficacy against viruses. Although TSB - based MS2 was shown to reduce free chlorine demand, we used MS2 TSB stock solutions in our inoculums for simulated commercial processing. Following inoculation, TSB - based cultures were diluted 3 - fold and further removal occurred following lettuce drying and shredding prior to flume washing. Therefore, the ef fect of TSB - cultured virus on free chlorine in the flume wash water suspens ion was expected to be minimal. However, this could represent the type of organic material that might be associated with naturally occurring viruses which are found on lettuce from fecal contamination. Processing and washing lettuce with water alone reduced MS2 populations on lettuce an average of 1 log (PFU/g). This is comparable to simulated commercial processin g and laboratory studies where virus reductions of ~0.5 to 1 log were reported for strawb erries and lettuce, respectively, following washing with tap water alone (Casteel et al., 2009, Fraisse et al., 2011) . In the present processing experiments , a 2 min wash in 25 ppm of free chlorine led to an 110 average re duction followin g processing of 0.8 log (PFU/g), which was not significantly different from pr ocessing with tap water (<0.05 ppm of free chlorine) alone. Hence, these manufacturer recommended free chlorine levels do not provide enhanced inactivation against non - enveloped , RNA viruses when present on fresh produce. However following processing, free chlorine levels were reduced by 5 to 10 ppm which , as already stated , reduces the efficacy of the disinfection process. It is likely that this is a result of the organic load introduced by the lettuce during fluming. Although chlorine levels were reduced, previous laboratory studies have show n that RNA virus reduction on fresh produce (strawberries, le ttuce, or tomato) ranged from 1 to 3 logs when exposed to 15 - 20 ppm of free chlorine for up to 10 minutes (Casteel et al., 2009; Fraisse et al., 2011) . In the present study, results were similar to a simulated commercial processing study where 20 ppm of free chlorine in wash water reduced MS2 populations on strawberries by ~1 lo g (Casteel et al., 2009) . Furthermore, laboratory studies using higher levels of free chlorine ran ging between 50 ppm and 800 ppm could not achieve more than a 3 log reduction of non - enveloped RNA viruses ( feline calicivirus, hepatitis A, MS2, murine noro virus ) when inoculated on fresh produce (Allwood et al., 2004, Dawson et al., 2005, Fraisse et al., 2011, Gulati et al. , 2001 ) . These data suggest that enteric viruses and enteric virus surrogates are very stable and resistant to current commercial chlori ne disinfection techniques used in food processing. Many studies investigating virus survival on fresh produce have also observed enhanced virus survival when attached to leafy greens and exposed to varying environmental conditions or disinfection techn iques (Badawy et al., 1985, Croci et al., 2002) . To date there is limited knowledge as to why virus s urvival is greater on lettuce compared to other food commodities. It is possible that enteric viruses are protected due to the rough or rigid surfaces of lettuce or resistant to naturally occurring antimicrobials (Badawy et al., 1985, Seymour & Appleton, 2001) . 111 In addition, a study where cut romaine lettuce was dip inoculated into biosolids containing murine norovirus has shown that viruses have the abili ty to internalize (access inner tissues) cut edges of leaflet surfaces which may offer protection from chemical disinfection (Wei et al., 2010) . The main reason for including a sanitizer in the wash water is to inactivate viruses after they have been re moved from fresh produce surfaces to ultimately prev ent cross contamination from viruses recirculating in the flume wash. Therefore, the flume water was s ampled following processing with the sanitizer. MS2 was present in al l flume water samples processed and was concentrated using ultrafiltration. Estimation of the reduction of MS2 in flume wash water suggested that 25 ppm free chlorine had greater efficacy in the wash water than on lettuce (~4 logs). However, the initial MS2 concentration in the wash w ater was calculated based on the difference in total MS2 ent ering and exiting the flume tank on lettuce and therefore our log reduction is an estimation, assuming all viruses were removed into the 800 L of wash water. These data together suggest that viru ses are very resistant to chlorine whe n present on the surface of romaine lettuce. In addition, since MS2 acts as a surrogate for foodborne viruses ( e.g. norovirus and hepatitis A), enteric viruses are also likely to persist on lettuce given current fresh produce processing practices . 112 V. CONCLUSIONS, LIMITAT IONS, AND FUTURE NEEDS Although i nnovative next - generation sequencing technologies have resulted in greater throughput and significant cost - reductions for sequencing entire microbial communities , there are still limitations when applied to viral metagenomics (Kircher & Kelso, 2010; L. Liu et al., 2012) . For example, there is currently no standard method for studying the virome in environmental samples. In addition, metagenomics must overcome bi as associated with each stage of virus recovery and detection including elution, concentration, purification, extraction, and PCR amplification (Aw et al., 2014 ) . Current computational tools used for bioinformatics analysis are intensive and the analysis of entire communities is difficult and time consuming. Novel bioinformatics tools are desir able for more rapid analysis of the large amount of data obtained from next - generation sequencing. Despite these limitations next - generation sequencing and metag enomics provides a broad view of food safety like never before, particularly for hazards such as viruses where traditional methods have limited the ability to monitor . Metagenomic techniques can be applied to a wide range of environmental samples includin g foods and the use and development of new genomics tools will aide in identifying and assessing better monitoring targets from harvesting to food processing . This is the first study to use next - generation sequencing and metagenomics to characterize the v irome in irrigation water and lettuce. The large number of no hits in each sample demonstrates the importance of building upon current viral databases to gain knowledge into the virus world. In addition, t hese results suggest that metagenomic technique s can be used to successfully identify specific viruses of importance to the food industry and public health, including foodborne human enteric viruses. Specifically, dsRNA enteric viruses including rotavirus A and human picobirnavirus were identified in irrigation water and lettuce, suggesting 113 these viruses are a food safety concern beginning at the farm - level stage of production. For future metagenomic studies on fresh produce, the author recommends a larger sample size at each stage of field production . In addition, in order to determine the viral risks from farm - to - fork, lettuce needs to be sampled throughout the supply chain including post - harvest production. Currently, viral metagenomic analysis of lettuce collected from distribution centers is bei ng conducted and will hopefully provide a better view of the lettuce virome and viral hazards following postharvest. Having found a wide variety of viruses , including human enteric viruses , on lettuce from the field using metagenomics there was interest in how removal might be achieved during processing after harvest. In this study, simulated commercial leafy green processing experiments demonstrated that MS2 is very resistant to chlorine when present on the surface of romaine lettuce. Since MS2 acts a s a surrogate for foodborne viruses , enteric viruses on lettuce are also likely to be very stable and resistant to current commercial chlorine disinfection techniques used with food processing systems . This suggests current recommended commercial productio n practices are unable to effectively decrease virus on leafy green during processing These data advocate that disinfection methods other than washing with a chlorine - based sanitizer should be investigated further for leafy green processing. To date, nu merous physical (ultrasound high pressure, ultraviolet, ionizing radiation) and chemical (chlorine dioxide, sodium chlorite, quaternary ammonium compounds, hydrogen peroxide, peroxyacetic acid) disinfection methods have been investigated for use in foods ( CFSAN , 2014) . Many viral disinfection studies on leafy greens in particular have investigated peroxyacetic acid and hydrogen peroxide as sanitizers. These chemicals have achieved 1 - 3 log reductions for feline calicivirus on leafy greens in laboratory stu dies. However, these same treatments can negatively impact end product 114 quality. The human and environmental safety of these compounds also warrants further investigation (Allwood et al., 2004, Baert et al., 2009, CFSAN , 2014, Fraisse et al., 2011, Gulati, et al. , 2001 ) . T his study demonstrates the importance of minimizing viral contamination on leafy greens prior to processing. There are many critical control points from farm - to - fork that aid in the reduction of fresh produce contamination. To date, t he FDA has developed laws, rules, and guidelines to combat human pathogen contamination and transmission in fresh produce, all of which recommend clean, safe water at all levels of production. At the pre - harvest level, potential sources of microbial conta mination include irrigation water, pesticide and fertilizer application, cooling, and frost control ( CFSAN , 2009) . In respect to irrigation water, recognition of the water source, historical use of the land, and existing human or agricultural practices a re all important considerations that help identify potential sources of microbial contamination ( CFSAN , 2009) . Recommended practices for limiting microbial hazards in irrigation water include restricting livestock and wildlife access to water to prevent f ecal contamination, well and septic tank maintenance, treating and testing, as well as conservation practices such as sod waterways, vegetative buffers, and runoff control structures that help limit runoff pollution ( CFSAN , 2009) . Worker hygiene is a majo r control point at both pre - and postharvest levels of production. Suggested practices to provide safe produce include proper hand washing, use of disposable gloves, and not coming to work when ill as well as employee training to address these expectation s. These practices combined with proper disinfection techniques are critical for supplying safe, quality leafy greens to the consumer and ultimately protecting public health and food safety. 115 APPENDIX 116 Figure A1 . Yuma irrigation water sampling locations 117 Figure A2 . Examples of Yuma furrow irrigation water (1), iceberg lettuce following harvest (2), and romaine lettuce following harvest (3) 118 Figure A3 . Ultrafiltration syste m for the primary concentration of irrigation water samples 119 Figure A4 . Lettuce centrifugal dryer (1) , shredder (2) , flume tank (3) , and mechanical shaker table (4) sampling locations 120 Table A1 . Raw contig sequence dat a used for Yuma irrigation water virome analysis Families Viral types Host YW1 YW2 YW3 YW4 YW5 YW6 Viruses Unclassified NA * 353 179 88 131 482 293 dsDNA viruses, no RNA stage dsDNA NA * 1993 397 534 852 2303 2038 Ascoviridae dsDNA Animals (invertebra te) 24 2 5 16 12 18 Asfarviridae dsDNA Animals (vertebrate) 2 0 0 0 0 0 Baculoviridae dsDNA Animals (invertebrate) 179 8 2 2 14 7 Bicaudaviridae dsDNA Archaea 2 0 0 0 0 0 Caudovirales dsDNA Bacteria 713 319 410 682 2106 1637 Myoviridae dsDNA Bacteria 1460 683 894 1498 3595 3400 Podoviridae dsDNA Bacteria 770 416 473 842 2488 1820 Siphoviridae dsDNA Bacteria 1111 489 607 1008 3019 2352 unclassified Caudovirales dsDNA Bacteria 53 31 38 69 251 166 Herpesvirales dsDNA Animals (vertebrate) 3 0 1 1 0 1 Alloherpesviridae dsDNA Animals (vertebrate) 17 2 3 3 10 10 Herpesviridae dsDNA Animals (vertebrate) 101 2 4 8 20 17 unclassified Herpesvirales dsDNA Animals (invertebrate) 2 0 1 5 2 4 Iridoviridae dsDNA Animals (vertebrate and invertebrate) 88 16 23 26 86 70 Ligamenvirales dsDNA Archaea 0 0 0 0 0 0 Lipothrixviridae dsDNA Archaea 2 2 0 3 5 3 Rudiviridae dsDNA Archaea 0 0 0 0 1 1 Marseilleviridae dsDNA Other (ameoba, protozoa, or fungi) 116 8 9 13 26 21 Mimiviridae dsDNA Other (ameoba, protozoa , or fungi) 1080 67 136 225 354 458 Nimaviridae dsDNA Animals (invertebrate) 7 0 0 0 2 0 Nudiviridae dsDNA Animals (invertebrate) 19 2 5 2 12 9 Papillomaviridae dsDNA Animals (vertebrate) 1 0 0 0 0 0 Phycodnaviridae dsDNA Algae 997 131 133 216 560 517 Polydnaviridae dsDNA Animals (invertebrate) 10 6 3 9 25 14 121 Polyomaviridae dsDNA Animals (vertebrate) 1 0 0 0 0 0 Poxviridae dsDNA Animals (vertebrate and invertebrate) 503 7 12 16 45 32 Tectiviridae dsDNA Bacteria 6 5 1 2 2 2 uncla ssified dsDNA phages Unclassified Bacteria 166 70 102 230 633 485 unclassified dsDNA viruses Unclassified NA * 485 35 63 108 235 240 dsRNA viruses dsRNA NA * 0 0 0 0 0 0 Chrysoviridae dsRNA Other (ameoba, protozoa, or fungi) 0 1 0 0 0 1 Cystoviridae dsRNA Bacteria 0 1 0 0 0 0 Endornaviridae dsRNA Plants 2 0 0 0 0 0 Partitiviridae dsRNA Other (ameoba, protozoa, or fungi) 4 4 2 2 0 5 Picobirnaviridae dsRNA Animals (vertebrate) 2 1 0 0 0 0 Reoviridae dsRNA Animals and plants 10 5 0 1 1 3 Totiviri dae dsRNA Other (ameoba, protozoa, or fungi) 3 1 0 0 0 0 unclassified dsRNA viruses Unclassified NA * 1 0 0 0 0 1 Retro - transcribing viruses Retro - transcribing NA * 1 0 1 0 1 0 Caulimoviridae Retro - transcribing Plants 1 0 0 0 1 2 Retroviridae Retro - tra nscribing Animals (vertebrate) 4 0 2 0 0 1 Satellites Satellites NA * 34 27 11 11 29 17 ssDNA viruses ssDNA NA * 306 250 64 38 301 109 Circoviridae ssDNA Animals (vertebrate and invertebrate) 350 304 75 58 257 147 Geminiviridae ssDNA Plants 136 113 21 11 90 41 Inoviridae ssDNA Bacteria 5 2 7 7 2 6 Microviridae ssDNA Bacteria 1471 1174 255 233 2900 440 Nanoviridae ssDNA Plants 44 37 11 5 23 19 Parvoviridae ssDNA Animals (vertebrate and invertebrate) 69 56 11 7 45 30 Parvovirus NIH - CQV ** NA * NA * 2 33 177 62 55 278 118 unclassified ssDNA viruses Unclassified NA * 602 475 92 71 672 184 122 ssRNA viruses ssRNA NA * 0 0 0 0 0 0 ssRNA negative - strand viruses ssRNA NA * 1 0 0 0 0 0 Bunyaviridae ssRNA Animals (vertebrate and invertebra te) 6 1 0 0 0 0 Tenuivirus (genus) ssRNA Plants 0 1 0 0 0 0 ssRNA positive - strand viruses, no DNA stage ssRNA NA * 44 38 7 5 50 9 Alphatetraviridae ssRNA Animals (invertebrate) 1 2 0 0 1 1 Astroviridae ssRNA Animals (vertebrate) 1 1 0 0 1 1 Benyvirid ae ssRNA Plants 6 5 6 3 4 4 Bromoviridae ssRNA Plants 1 1 0 2 1 1 Caliciviridae ssRNA Animals (vertebrate) 2 1 0 0 1 0 Carmotetraviridae ssRNA Animals (invertebrate) 0 0 0 0 1 1 Closteroviridae ssRNA Plants 1 1 0 0 1 0 Hepeviridae ssRNA Animals (ve rtebrate) 6 10 1 0 7 5 Higrevirus (genus) ssRNA Plants 1 0 0 0 0 0 Leviviridae ssRNA Bacteria 30 17 9 13 0 1 Luteoviridae ssRNA Plants 2 0 0 0 1 0 Narnaviridae ssRNA Other (ameoba, protozoa, or fungi) 0 1 0 0 0 0 Nidovirales ssRNA Animals (vertebra te and invertebrate) 1 0 0 0 0 0 Coronaviridae ssRNA Animals (vertebrate) 4 0 0 0 0 0 Roniviridae ssRNA Animals (invertebrate) 2 0 0 0 0 0 Nodaviridae ssRNA Animals (vertebrate) 14 17 0 0 15 0 Ourmiavirus (genus) ssRNA Plants 38 32 4 7 27 7 Picornavi rales ssRNA Animals and Plants 115 94 16 33 112 27 Dicistroviridae ssRNA Animals (invertebrate) 198 188 60 56 234 65 environmental samples < Picornavirales > ssRNA NA * 228 201 39 81 269 82 Iflaviridae ssRNA Animals (invertebrate) 7 5 0 0 3 1 123 Table A1 ( Marnaviridae ssRNA Plants 4 3 2 2 6 2 Picornaviridae ssRNA Animals (vertebrate) 10 12 0 3 8 3 Secoviridae ssRNA Plants 4 7 2 1 3 1 unassigned Picornavirales ssRNA NA * 23 20 8 5 22 6 unclassified Picornavirales ssRNA Animals (vertebrate) 2 0 0 0 0 0 Potyviridae ssRNA Plants 9 5 1 3 7 5 Sobemovirus (genus) ssRNA Plants 10 7 1 1 3 3 Tombusviridae ssRNA Plants 99 90 19 23 74 27 Tymovirales ssRNA Plants 0 0 0 0 0 0 Tymoviridae ssRNA Plants 1 0 0 0 0 0 Umbravirus (genus) ssRNA Plants 1 1 0 0 1 0 unclassified ssRNA positive - strand viruses ssRNA NA * 102 86 18 32 117 47 Virgaviridae ssRNA Plants 4 2 0 0 0 0 unassigned ssRNA viruses Unclassified NA * 0 0 0 0 0 0 Alvernaviridae ssRNA Algae 4 1 0 2 2 1 unassigned viruses Unclassified NA * 0 0 0 0 0 0 Bacilladnavirus Unclassified Algae 3 2 2 1 4 5 Bidnaviridae Unclassified Animals (invertebrate) 2 3 1 1 1 2 Hytrosaviridae Unclassified Animals (invertebrate) 9 1 0 1 3 0 unclassified phages Unclassified Bacteria 19 9 10 14 50 40 uncla ssified virophages Unclassified NA * 10 1 2 11 37 20 unclassified viruses Unclassified NA * 3 2 1 0 4 3 Not assigned NA * NA * 5491 1477 1144 1791 5695 4145 No hits NA * NA * 107829 14533 13022 19023 49695 40505 * NA. Not applicable, data was no t used in category analysis ** Contaminant removed from analysis 124 Table A2 . Raw contig sequence data used for Yuma lettuce virome analysis Viral Assignment Viral Type Host Iceberg Control Iceberg Worker Harvest Iceberg Post Worker Break Romaine Control Romaine Wor ker Harvest Romaine Chop and Wash Romaine Mixed Salad Viruses Unclassified NA* 47 17 11 49 14 3 8 dsDNA viruses, no RNA stage dsDNA NA* 152 94 74 93 66 33 76 Ascoviridae dsDNA Animals (invertebrate) 4 1 2 0 0 1 1 Asfarviridae dsDNA Animals (vertebrate) 0 0 0 0 1 0 0 Baculoviridae dsDNA Animals (invertebrate) 2 3 1 1 1 0 1 Caudovirales dsDNA Bacteria 137 34 27 159 31 9 33 Myoviridae dsDNA Bacteria 205 150 78 400 71 12 56 Podoviridae dsDNA Bacteria 175 64 42 188 51 11 53 Siphoviridae dsDNA Bacteria 3 40 134 80 427 90 15 78 unclassified Caudovirales dsDNA Bacteria 9 2 0 2 2 0 4 Herpesvirales dsDNA Animals (vertebrate) 0 0 0 0 1 0 1 Alloherpesviridae dsDNA Animals (vertebrate) 0 0 1 0 0 0 2 Herpesviridae dsDNA Animals (vertebrate) 3 2 2 2 0 0 1 uncl assified Herpesvirales dsDNA Animals (vertebrate) 0 1 1 0 0 0 0 Iridoviridae dsDNA Animals (vertebrate and invertebrate) 27 27 14 22 12 15 12 Marseilleviridae dsDNA Other (ameoba, protozoa, or fungi) 2 3 4 1 1 2 3 Mimiviridae dsDNA Other (ameoba, protoz oa, or fungi) 68 53 36 38 25 12 31 125 Nudiviridae dsDNA Animals (invertebrate) 0 0 0 0 1 0 0 Papillomaviridae dsDNA Animals (vertebrate) 1 0 0 0 0 0 0 Phycodnaviridae dsDNA Algae 74 40 27 29 26 13 31 Polydnaviridae dsDNA Animals (inverte brate) 4 0 2 3 2 0 0 Poxviridae dsDNA Anima l s (vertebrate and invertebrate) 6 5 5 4 4 1 1 Salterprovirus (genus) dsDNA Archaea 1 0 0 0 0 0 1 Tectiviridae dsDNA Bacteria 5 0 1 1 0 0 1 unclassified dsDNA phages Unclassified NA* 38 15 4 56 6 2 17 unclass ified dsDNA viruses Unclassified NA* 15 10 4 10 8 0 5 dsRNA viruses dsRNA NA* 0 1 0 1 3 1 0 Birnaviridae dsRNA Animals (vertebrate and invertebrate) 1 0 0 0 0 0 0 Chrysoviridae dsRNA Other (ameoba, protozoa, or fungi) 0 2 0 4 1 0 1 Cystoviridae dsRNA B acteria 2 0 0 0 0 0 0 Endornaviridae dsRNA Plants 9 45 25 74 63 19 52 Hypoviridae dsRNA Other (ameoba, protozoa, or fungi) 0 0 0 1 0 0 0 Partitiviridae dsRNA Other (ameoba, protozoa, or fungi) 18 16 7 79 48 25 26 Picobirnaviridae dsRNA Animals (vertebr ate) 1 0 0 0 4 1 1 Reoviridae dsRNA Animals and plants 0 26 10 3 23 4 16 126 Totiviridae dsRNA Other (ameoba, protozoa, or fungi) 28 51 9 88 95 48 64 unclassified dsRNA viruses Unclassified NA* 15 31 21 31 23 21 22 Retro - transcribing viru ses Retro - transcribing NA* 30 50 32 25 31 42 26 Caulimoviridae Retro - transcribing Plants 238 357 200 189 168 157 158 Retroviridae Retro - transcribing Animals (vertebrate) 103 193 137 61 99 130 98 Satellites Satellites NA* 0 0 0 2 0 0 0 ssDNA viruses ssD NA NA* 17 5 2 6 3 2 2 Circoviridae ssDNA Animals (vertebrate and invertebrate) 12 5 2 10 3 2 3 Geminiviridae ssDNA Plants 2 1 0 0 1 1 0 Inoviridae ssDNA Bacteria 43 25 21 27 18 10 11 Microviridae ssDNA Bacteria 56 15 4 12 13 2 5 Nanoviridae ssDNA Plan ts 0 0 0 0 0 0 1 Parvoviridae ssDNA Animals (vertebrate and invertebrate) 0 2 0 4 0 1 0 Parvovirus NIH - CQV** NA* NA* 2 1 0 2 0 0 2 unclassified ssDNA viruses Unclassified NA* 43 17 4 54 25 11 7 ssRNA viruses ssRNA NA* 0 0 0 0 0 0 0 ssRNA negative - stra nd viruses ssRNA NA* 0 5 4 1 3 1 1 Bunyaviridae ssRNA Animals (vertebrate and invertebrate) 0 2 1 0 1 1 3 Mononegavirales ssRNA Animals and plants 1 0 0 0 0 1 0 127 Rhabdoviridae ssRNA Animals and plants 0 0 0 0 0 0 1 Ophioviridae ssRNA P lants 0 0 0 2 1 0 0 Tenuivirus (genus) ssRNA Plants 3 2 2 1 4 1 1 Varicosavirus ssRNA Plants 50 26 12 3 1 0 0 ssRNA positive - strand viruses, no DNA stage ssRNA NA* 3 5 0 6 4 3 1 Bromoviridae ssRNA Plants 0 3 0 2 4 0 3 Cilevirus (genus) ssRNA Plants 0 0 0 1 0 0 0 Closteroviridae ssRNA Plants 87 69 28 120 65 31 53 Flaviviridae ssRNA Animals (vertebrate and invertebrate) 1 1 0 0 1 1 0 Hepeviridae ssRNA Animals (vertebrate) 0 0 0 2 0 0 0 Leviviridae ssRNA Bacteria 2 0 0 1 0 0 0 Narnaviridae ssRNA Othe r (ameoba, protozoa, or fungi) 1 3 1 5 9 3 3 Nodaviridae ssRNA Animals (vertebrate) 0 0 0 4 2 0 3 Ourmiavirus (genus) ssRNA Plants 5 2 1 2 4 0 0 Picornavirales ssRNA Animals and Plants 3 2 2 1 0 0 0 Dicistroviridae ssRNA Animals (invertebrate) 7 10 0 11 20 1 7 environmental samples < Picornavirales > ssRNA NA* 1 0 0 1 0 0 0 Iflaviridae ssRNA Animals (invertebrate) 10 7 3 23 2 7 0 Picornaviridae ssRNA Animals (vertebrate) 0 0 0 1 0 0 0 128 unclassified Picornavirales ssRNA Animals (ver tebrate) 0 0 0 1 1 1 1 Potyviridae ssRNA Plants 0 2 1 2 3 1 1 Tombusviridae ssRNA Plants 3 0 0 3 7 1 0 Tymovirales ssRNA Plants 0 0 0 0 0 0 0 Betaflexiviridae ssRNA Plants 0 1 0 0 0 0 0 Tymoviridae ssRNA Plants 1 0 0 0 0 0 0 Umbravirus (genus) ssRNA Plants 0 0 0 0 2 2 0 unclassified ssRNA positive - strand viruses Unclassified NA* 2 0 0 4 1 0 0 Virgaviridae ssRNA Plants 0 0 0 0 1 0 2 unassigned viruses Unclassified NA* 0 0 0 0 0 0 0 Bacilladnavirus (genus) Unclassified Other (ameoba, protozoa, or fu ngi) 2 2 0 0 1 0 0 Bidnaviridae Unclassified Animals (invertebrate) 0 0 0 5 0 0 0 unclassified phages Unclassified Bacteria 5 5 8 8 4 1 5 Not assigned NA* NA* 388 252 179 239 148 96 116 No hits NA* NA* 17004 12563 10288 10611 7298 6962 6884 * NA. Not applicable, data was no t used in category analysis ** Contaminant removed from analysis 129 * Sample ID descriptions: lettuce sampled before processing (BP) and after shredding (S R), flume washing (F), shaking (SA), and centrifugal drying (C). Centrifuge water (CW) and the Inoculated concentrate (IC) are also included. 1, 2 and 3 are triplicate samples * * Calculated using equations in Part III, Section 3.1.3, pages 39 - 40 ** * NR no result, data were unavailable due no plaques in countable range (dilution error) ****NA not applicable Table A3 . Raw data and the calculated MS2 eluent plaque concentration, lettuce concentration, and average lettuce concentration for samples collected during trial 1 leafy green processing without sanitizer Sample ID * Lettuce weight (g) Eluent volume (mL) Dilution Plate 1 Plate 2 Plate 3 Eluent plaque concentration (PFU/mL) ** Phage concentration on lettuce (PFU/g) ** Average phage concentration on l ettuce for each s ample location (PFU/g) BP 1 50.00 234.50 10 - 4 108 101 92 5.02x10 5 2.35x10 6 1.38x10 6 BP 2 50.00 240.00 10 - 4 49 37 43 2.15x10 5 1.03x10 6 BP 3 50.00 233.00 10 - 4 33 25 40 1.63 x10 5 7.61x10 5 SR 1 52.10 235.00 10 - 4 20 26 29 1.25x10 5 5.64 x10 5 5.97 x10 5 SR 2 54.20 235.00 10 - 4 38 37 20 1.5 8x10 5 6.87x10 5 SR 3 59.30 235.00 10 - 4 38 34 10 1.37x10 5 5.42x10 5 F1 58.00 235.00 10 - 3 91 64 82 3.95x10 4 1.60x10 5 1.64x10 5 F2 51.20 240.00 10 - 3 37 75 66 2.97x10 4 1.39x10 5 F3 53.30 235.00 10 - 3 85 114 63 4.37x10 4 1.93x10 5 SA 1 5 0.30 240.00 10 - 3 13 14 55 1.37x10 4 6.52x10 4 7.61x10 4 SA 2 57.20 235.00 10 - 3 40 22 20 1.37x10 4 5.61x10 4 SA 3 53.20 237.00 10 - 3 45 51 48 2.40x10 4 1.07x10 5 C1 51.60 240.00 10 - 3 NR*** NR*** NR*** NR*** NR*** 1.11x10 5 C2 51.20 240.00 10 - 3 X X 53 2.65 x10 4 1 .24 x10 5 C3 55.40 237.00 10 - 3 40 42 55 2.28x10 4 9.77 x10 4 CW NR*** 500.00 10 - 4 64 52 55 2.85 x10 5 NA**** NA**** IC NR*** NR*** 10 - 9 65 48 53 2.77 x10 10 NA**** NA**** 130 Table A4 . Raw Data and the calculated MS2 eluent plaque concentration, lettuce concentration, and average lettuce concentration for samples collecte d during trial 2 leafy green processing without sanitizer * Sample ID * Lettuce Weight (g) Eluent Volume (mL) Dilution Plate 1 Plate 2 Plate 3 Eluent plaque concentration (PFU/mL) * * Phage concentration on lettuce (PFU/g) * * Average Phage concentration on Le ttuce (PFU/g) BP1 49.9 241.0 10 - 4 30 38 29 1.62x10 5 7.81x10 5 7.81x10 5 SR1 50.8 238.0 10 - 4 20 40 23 1.38x10 5 6.48x10 5 6.81x10 5 SR2 50.7 237.0 10 - 4 21 26 33 1.33x10 5 6.23x10 5 SR3 50.5 236.0 10 - 4 34 23 42 1.65x10 5 7.71x10 5 F1 50.2 250.0 10 - 3 30 34 34 1 .63 x10 4 8.13x10 4 7.07x10 4 F2 50.2 249.0 10 - 3 27 30 40 1.62x10 4 8.02x10 4 F3 50.1 205.0 10 - 3 22 23 29 1.23x10 4 5.05x10 4 SA1 52.0 248.0 10 - 3 13 34 24 1.18x10 4 5.64x10 4 5.39x10 4 10 - 2 167 205 191 9.38x10 3 4.48x10 4 SA2 52.1 249.0 10 - 3 25 27 20 1.20x10 4 5.74x10 4 10 - 2 242 257 160 1.10x10 4 5.25x10 4 SA3 54.6 250.0 10 - 3 34 33 19 1.43x10 4 6.56x10 4 10 - 2 203 182 229 1.02x10 4 4.69x10 4 C1 53.7 240.0 10 - 3 34 45 28 1.40x10 4 6.26x10 4 9.48x10 4 C2 53.4 241.0 10 - 3 56 36 50 2.50x10 4 1.13x10 5 C3 55.5 24 0.0 10 - 3 47 36 53 2.27x10 4 9.80x10 4 CW NA *** 500.00 10 - 4 37 33 21 1.52x10 5 NA*** NA*** I C NA*** NA*** 10 - 9 54 43 38 2.25x10 10 NA*** NA*** * Sample ID descriptions: lettuce sampled before processing (BP) and after shredding (SR), flume washing (F), shaki ng (SA), and centrifugal drying (C). Centrifuge water (CW) and the Inoculated concentrate (IC) are also included. 1, 2 and 3 are triplicate samples. * * Calculated using equations in Part III, Section 3.1.3, pages 39 - 40 ** *NA not applicable 131 Table A5 . Raw data and the calculated MS2 eluent plaque concentration, lettuce concentration, and average lettuce concentration for samples collected during trial 3 leafy green processing without sanitizer Sample ID * Lettuce Weight (g) Eluent Volume (mL) Dilution P late 1 Plate 2 Plate 3 Eluent plaque concentration (PFU/mL) * * Phage concentration on lettuce (PFU/g) * * Average Phage concentration on Lettuce (PFU/g) BP1 49.9 238.0 10 - 4 39 25 48 1.87x10 5 8.90x10 5 7.12x10 5 BP2 54.6 236.0 10 - 4 21 30 20 1.18x10 5 5.11x10 5 BP3 50.3 236.0 10 - 4 30 32 32 1.57x10 5 7.35x10 5 SR1 51.1 237.0 10 - 4 22 14 28 1.07x10 5 4.94x10 5 5.20x10 5 10 - 3 219 275 224 1.20x10 5 5.55x10 5 SR2 51.6 235.0 10 - 4 21 28 25 1.23x10 5 5.62x10 5 10 - 3 224 211 230 1.11x10 5 5.05x10 5 SR3 52.0 235.0 10 - 4 24 24 30 1.30x10 5 5.88x10 5 10 - 3 205 192 154 9.18x10 4 4.15x10 5 F1 53.4 241.0 10 - 3 29 26 26 1.35x10 4 6.10x10 4 5.80x10 4 F2 53.0 242.0 10 - 3 28 21 24 1.22x10 4 5.55x10 4 F3 53.2 245.0 10 - 3 31 23 21 1.25x10 4 5.76x10 4 SA1 50.3 240.0 10 - 2 305 311 314 1.5 5x10 4 7.39x10 4 7.38x10 4 SA2 51.4 220.0 10 - 2 296 344 265 1.51x10 4 6.46x10 4 SA3 51.1 240.0 10 - 2 386 318 354 1.76x10 4 8.29 x10 4 C1 52.2 235.0 10 - 2 381 400 329 1.85x10 4 8.33x10 4 8.66x10 4 C2 50.2 235.0 10 - 2 461 406 371 2.06x10 4 9.67 x10 4 C3 51.0 235.0 10 - 2 371 345 324 1.73x10 4 7.99x10 4 CW NA*** 500.00 10 - 4 22 21 18 1.02 x10 5 NA*** NA*** 10 - 3 186 208 200 9.90 x10 4 NA*** NA*** I C NA*** NA*** 10 - 9 30 23 26 1.32 x10 10 NA*** NA*** * Sample ID descriptions: lettuce sampled before processing (BP) and after shredding (SR), flume washing (F), shaking (SA), and centrifugal drying (C). Centrifuge water (CW) and the Inoculated concentrate (IC) are also included. 1, 2 and 3 are triplicate samples. * * Calculated using equations in Part III, Section 3.1.3, pages 39 - 40 ***NA not applicable 132 * Sample ID descriptions: lettuc e sampled before processing (BP) and after shredding (SR), flume washing (F), shaking (SA), and centrifugal drying (C). Centrifuge water (CW) and the Inoculated concentrate (IC) are also included. 1, 2 and 3 are triplicate samples. * * Calculated using e quations in Part III, Section 3.1.3, pages 39 - 40 ***NA not applicable Table A6 . Raw data and the calculated MS2 eluent plaque concentration, lettuce concentration, and average lettuce concentration for samples collected during trial 1 leafy green processing with 25 ppm free chlorine * Sampl e ID * Lettuce Weight (g) Eluent Volume (mL) Dilution Plate 1 Plate 2 Plate 3 Eluent plaque concentration (PFU/mL) * * Phage concentration on lettuce (PFU/g) * * Average Phage concentration on Lettuce (PFU/g) Control 51.4 200.0 10 0 2 2 5 1.50x10 0 5.84 x10 0 3.51 x10 0 Control 49.8 235.0 10 0 1 0 0 5.00x10 - 1 2.36 x10 0 Control 50.6 235.0 10 0 1 0 1 5. 00x10 - 1 2.32 x10 0 BP1 50.2 233.0 10 - 3 37 44 33 1.90x10 4 8.82x10 4 1.20x10 5 BP2 50.7 226.0 10 - 3 46 58 53 2.62x10 4 1.17x10 5 BP3 49.5 236.0 10 - 3 69 59 66 3.23x10 4 1.54x1 0 5 SR1 49.9 238.0 10 - 3 41 39 20 1. 67x10 4 7.95x10 4 7.19x10 4 SR2 51.2 237.0 10 - 3 30 26 33 1.48x10 4 6.87x10 4 SR3 49.3 238.0 10 - 3 24 27 33 1.40x10 4 6.76x10 4 F1 49.1 249.0 10 - 2 101 91 109 5.02 x10 3 2.54x10 4 2.56 x10 4 F2 49.9 250.0 10 - 2 85 99 97 4.68x10 3 2 .35x10 4 F3 50.0 250.0 10 - 2 134 102 100 5.60x10 3 2.80x10 4 SA1 50.9 250.0 10 - 2 80 81 106 4.45x10 3 2.19x10 4 2.26x10 4 SA2 49.6 246.0 10 - 2 75 70 87 3.87x10 3 1.92x10 4 SA3 49.8 247.0 10 - 2 109 114 102 5.42x10 3 2.69x10 4 C1 50.0 245.0 10 - 2 101 119 129 5.82x 10 3 2.85x10 4 2.81x10 4 C2 50.0 245.0 10 - 2 119 113 120 5.87x10 3 2.87x10 4 C3 50.4 243.0 10 - 2 106 107 124 5.62x10 3 2.71x10 4 CW NA*** 500 10 - 3 33 32 38 1.72x10 4 NA*** NA*** I C NA*** NA*** 10 - 8 63 64 59 3.10x10 9 NA*** NA*** 133 * Sample ID descriptions: lettuce sampled before processing (BP) and after shredding (SR), flume washing (F), shaking (SA ), and centrifugal drying (C). Centrifuge water (CW) and the Inoculated concentrate (IC) are also included. 1, 2 and 3 are triplicate samples. * * Calculated using equations in Part III, Section 3.1.3, pages 39 - 40 ***NA not applicable Table A7 . Raw data and the calculated MS2 eluent plaque concentration, lettuce concentration, and average lettuce concentration for samples collected during trial 2 leafy green proc essing with 25 ppm free chlorine Sample ID * Lettuce Weight (g) Eluent Volume (mL) Dilution Plate 1 Plate 2 Plate 3 Eluent plaque concentration (PFU/mL) * * Phage concentration on lettuce (PFU/g) * * Average Phage concentration on Lettuce (PFU/g) BP1 49.8 2 35.0 10 - 3 27 34 42 1.72x10 4 8.10x10 4 7.70x10 4 BP2 50.0 233.5 10 - 3 33 38 47 1.97x10 4 9.18x10 4 BP3 50.7 226.5 10 - 3 23 29 26 1.30x10 4 5.81x10 4 SR1 50.1 236.0 10 - 3 45 40 34 1.98x10 4 9.34x10 4 8.87 x10 4 SR2 50.0 235.0 10 - 3 29 28 58 1 .92x10 4 9.01x10 4 SR3 4 9.8 235.0 10 - 3 34 28 43 1.75x10 4 8.26x10 4 F1 50.0 244.0 10 - 2 35 45 59 2.32x10 3 1.13x10 4 8.49x10 3 F2 50.9 238.0 10 - 2 30 24 35 1.48 x10 3 6.94x10 3 F3 49.8 245.0 10 - 2 23 28 37 1.47x10 3 7.22x10 3 SA1 50.1 246.0 10 - 2 31 45 42 1.97x10 3 9.66x10 3 8.36x10 3 SA 2 49.9 246.0 10 - 2 32 30 29 1.52x10 3 7.48x10 3 SA3 50.1 244.0 10 - 2 24 40 34 1.63x10 3 7.95x10 3 C1 50.0 226.0 10 - 2 36 30 34 1.67x10 3 7.53x10 3 8.96x10 3 C2 50.1 237.0 10 - 2 36 51 47 2.23x10 3 1.06x10 4 C3 50.5 211.0 10 - 2 53 39 34 2.10x10 3 8.77x10 3 CW NA*** 500.00 10 - 3 9 17 19 9.00x10 3 NA*** NA*** 10 - 2 171 130 187 8.13x10 3 NA*** NA*** I C NA*** NA*** 10 - 8 54 82 49 3.08x10 9 NA*** NA*** 134 * Sample ID descriptions: lettuce sampled before processing (BP) and after shredding (SR), flume washing (F), shaking (SA), and centrifugal drying (C). Centrifuge water (CW) and the Inoculated concentrate (IC) are also included. 1, 2 and 3 are triplicate samples. * * Calculated using equations in Part III, Section 3.1.3, pages 39 - 40 ***NA not applicable Table A8 . Raw data and the calculated MS2 eluent plaque concentration, lettuce concentration, and average lettuce concentration for samples collected during trial 3 leafy green processing with 25 ppm free chlorine Sample ID * Lettuce Weight (g) Eluent Volume (mL) Dilut ion Plate 1 Plate 2 Plate 3 Eluent plaque concentration (PFU/mL) * * Phage concentration on lettuce (PFU/g) * * Average Phage concentration on Lettuce (PFU/g) BP1 50.3 235.0 10 - 3 42 44 42 2.13x10 4 9.96 x10 4 6.64x10 4 BP2 50.6 235.0 10 - 3 8 32 28 1.50 x10 4 6. 97 x10 4 10 - 2 228 224 275 1.21x10 4 5.63x10 4 BP3 49.9 235.0 10 - 3 20 20 28 1.13x10 4 5.33 x10 4 10 - 2 205 210 261 1.13x10 4 5 .30x10 4 SR1 50.0 236.0 10 - 3 30 39 41 1.83x10 4 8.65x10 4 8.77 x10 4 SR2 50.1 239.0 10 - 3 47 34 25 1.77x10 4 8.43x10 4 SR3 50.1 235 .0 10 - 3 39 34 45 1.97x10 4 9.23x10 4 F1 50.1 250.0 10 - 2 53 44 43 2.33x10 3 1.16x10 4 1.48 x10 4 F2 50.4 250.0 10 - 2 55 58 56 2.82x10 3 1.40x10 4 F3 50.6 250.0 10 - 2 60 58 111 3.82x10 3 1.89x10 4 SA1 50.0 250.0 10 - 2 50 66 67 3.05x10 3 1.52x10 4 1.13 x10 4 SA2 50.2 250.0 10 - 2 31 39 36 1.77x10 3 8.79 x10 4 SA3 50.3 250.0 10 - 2 34 43 43 2.00x10 3 9.94x10 3 C1 50.0 243.0 10 - 2 36 53 38 2.12x10 3 1.03 x10 4 1.01 x10 4 C2 50.0 243.0 10 - 2 28 28 51 1.78x10 3 8.66x10 3 C3 50.0 243.0 10 - 2 35 59 46 2.33 x10 3 1.13x10 4 CW NA*** 720.00 10 - 2 94 109 92 4.92 x10 3 NA*** NA*** 10 - 3 14 20 12 7.67x10 3 NA*** NA*** I C NA*** NA*** 10 - 8 94 109 92 4.92x10 9 NA*** NA*** 135 Table A9. Flume water r aw data and MS2 concentration for sanitizer experiments Tria l Sampl e ID Concentrate volume (mL) Dilutio n Plate 1 Plate 2 Plate 3 Plate 4 Plate 5 Plate 6 Plate 7 Plate 8 PFU/m L Average PFU/mL 1 FW1 350 0 2 2 6 6 8 8 6 5 2.7 2.3 FW2 350 0 5 3 4 4 1 7 4 2 1.9 2 FW1 330 0 11 12 8 9 6 NR NR NR 4.6 3.8 FW2 250 0 4 6 7 4 8 NR NR NR 2.9 3 FW1 250 0 5 2 6 5 0 NR NR NR 2.3 3.4 FW2 230 0 9 5 10 10 12 NR NR NR 4.6 *NR no result 136 REFERENCES 137 REFERENCES Ackermann, H. - W. (2009). P hage classification and characterization. Methods in Molecular Biology (Clifton, N.J.) , 501 , 127 40. Adams, I. P., Glover, R. H., Monger, W. A ., Mumford, R., Jackeviciene, E., Navalinskiene, M., Samuitiene, M., & Boonham, N. (2009). Next - generation sequen cing and metagenomic analysis: A universal diagnostic tool in plant virology. Molecular Plant Pathology , 10 (4), 537 545. Allwood, P. B., Malik, Y. S., Hedberg, C. W., & Goyal, S. M. (2004). Effect of temperature and sanitizers on the survival of feline ca licivirus, Escherichia coli , and F - specific coliphage MS2 on leafy salad vegetables. Journal of Food Protection , 67 (7), 1451 6. Alum, A., Enriquez, C., & Gerba, C. P. (2011). Impact of drip irrigation method, s oil, and virus type on tomato and cucumber c on tamination. Food and Environmental Virology , 3 (2), 78 85. Angly, F. E., Felts, B., Breitbart, M., Salamon, P., Edwards, R. A., Carlson, C., Chan, A. M., Haynes, M., Kelley, S., Liu, H., Mahaffy, J. M., Mueller, J. E., Nulton, J., Olson, R., Parsons, R., R ayhawk, S., Suttle, C. A., & Rohwer, F. (2006). The marine viromes of four oceanic regions. PLoS Biology , 4 (11), e368. Asensio, L., González, I., García, T., & Martín, R. (2008). Determination of food authenticity by enzyme - linked immunosorbent assay (ELIS A). Food Control , 19 (1), 1 8. Aw, T. G., & Rose, J. B. (2012). Detection of pathogens in water: from phylochips to qPCR to pyrosequencing. Current Opinion in Biotechnology , 23 (3), 422 30. Aw, T. G. , Howe, A., & Rose, J. B. (2014 ). Metagenomic approaches for direct and cell culture evaluation of the virological quality of wastewater. Journal of Virological Methods , 210 , 15 21. Babraham Bioinformatics. ( 2015 ). FastQC a quality control tool for high throughput sequence d ata . Retrieved from http://www.bioinf ormatics.babraham.ac.uk/projects/fastqc/ . Accessed on 3 August, 2015. Badawy, A. S., Gerba, C. P., & Kelley, L. M. (1985). Survival of rotavirus SA - 11 on vegetables. Food Microbiology , 2 (3), 199 205. Baert, L., Debevere, J., & Uyttendaele, M. (2009). The effi cacy of preservation methods to inactivate foodborne viruses. International Journal of Food Microbiology , 131 (2 ), 83 94. 138 Bartlett, J. M. S., & Stirling, D. (2003). A short history of th e polymerase chain reaction. Methods in molecular biology (Clifton, N. J.) ( 3 - 6). Humana Press. Baudoux, A. - C., & Brussaard, C. P. D. (2005). Characterization of different viruses infecting the marine harmful algal bloom species Phaeocystis globosa . Virology , 341 (1), 80 90. Bergholz, T. M., Moreno Switt, A. I., & Wiedmann, M. (2014). Omics approaches in food safety: fulfilling the promise? Trends in Microbiology , 22 (5), 275 281. Beuchat, L. R. (2002). Ecological factors influencing survival and growth of human pathogens on raw fruits and vegetables. Microbes and Infection , 4 (4), 413 423. Bibby, K., Viau, E., & Peccia, J. (2011). Viral metagenome analysis to guide human pathogen monitoring in environmental samples. Letters in Applied Microbiology , 52 (4), 386 92. Bidawid, S ., Farber, J. M., & Sattar, S. A . (2000). Rapid conc entration and detection of hepatitis A virus from lettuce and strawberries. Journal of Virological Methods , 88 (2), 175 85. Biopieces. ( 2015 ) . Biopieces. Retrieved from https://code.google.com/p/biopieces/ . Accessed on 3 August 2015 . Bolger, A. M., Lohse, M., & Usadel, B. (2014). Trimmomatic: a flexible trimmer for Illumina sequence data. Bioin formatics (Oxford University Press ) , btu 170 . Boulila, M. (2011). Positive selection, molecular recombination structure and phylogenetic reconstruction of members of the family Tombusviridae : i mplication in virus taxonomy. Genetics and Molecular Biology , 34 (4), 647 60. Breitbart, M., & Rohwer, F. (2005). Here a virus, there a virus, everywhere the same virus? Trends in Microbiology , 13 (6), 278 84. Breitbart, M., Salamon, P., Andresen, B., Mahaffy, J. M., Segall, A. M., Mead, D., Azam, F., & Rohwer, F. (2002). Genomic analysis of uncultured marine viral communities. Proceedings of the National Academy of Sciences of the United States of America , 99 (22), 14250 14255. Brus saard, C. (2004). Viral control of p hyto plankton populations - a r eview. The Journal of Eukaryotic Microbiology , 51 (2), 125 138. Buchholz, A. L., Davidson, G. R., Marks, B. P., Todd, E. C. D., & Ryser, E. T. (2012 a). Quantitative t ransfer of Escherichia col i O157:H7 to equipment during small - scale production of fresh - cut leafy g reens. Journal of Food Protection , 75 (7), 1184 1197. 139 Buchholz, A. L., Davidson, G. R., Marks, B. P., Todd, E. C. D., & Ryser, E. T. (2012b). Transfer of Escherichia coli O157:H7 from equipment surfaces to fresh - cut leafy greens during processing in a model pilot - plant production line with sanitizer - free water. Journal of Food Protection , 75 (11), 1920 9. Butot, S., Putallaz, T., & Sánchez, G. (2007). Procedure for rapid concentration and detection of enteric viruses from berries and vegetables. Applied and Environmental Microbiology , 73 (1), 186 92. Calder, L., Simmons, G., Thornley, C., Taylor, P., Pritchard, K., Greening, G., & Bishop, J. (2003). An outbreak of hepatitis A associated with consumption of raw blueberries. Epidemiology and Infection , 131 (1), 745 - 51. Casteel, M. J., Schmidt, C. E., & Sobsey, M. D. (2008). Chlorine disinfection of produce to inactivate hepatitis A virus and coliphage MS2. International Journal of Food Mic robiology , 125 (3), 267 273. Casteel, M. J., Schmidt, C. E., & Sobsey, M. D. (2009). Chlorine inactivation of coliphage MS2 on strawberries by industrial - scale water washing units. Journal of Water and Health , 7 (2), 244 250. Center for Agricultural and R ur al Development (CARD). ( 2004 ) . The expanding U.S. market for fresh p roduce . Retrieved from http://www.agmrc.org/media/cms/ expandingusmarketproduce _07C66B1 CA5F70.pdf . Accessed on June 26, 2014 . Centers for Disease Control and Prevention (CDC) . (2011) . I ncidence and trends of infection with pathogens transmitted commonly through food foodborne diseases active surveillance network, 10 U.S. sites, 1996 - 2012. Morbidity and Mortality Weekly Report, 62 (15), 283 - 287. Centers for Dise ase Control and Prevention (CDC) . ( 2013 a) . Global foodborne infections network, formerly k nown as WHO global salm - s urv (GFN). Retrieved from http://www.cdc.gov/ncezid/dfwed/international/gfin.html . Accessed on June 22 2015. Centers for Disease Control and Prevention (CDC) . ( 2013 b) . Reporting a nd surveillance for norovirus. Retrieved from http://www.cdc.gov/norovirus/ reporting/index.html. Accessed on 26 March 2015. Centers for Disease Control and Prevention (CDC). ( 2014 ) . 2011 - Estimates of foodborne i llness in the United States - CDC 2011 e stim ates. Retrieved from http://www.cdc.gov/ foodborneburden/2011 - foodborne - estimates.html . Accessed on 23 June 2014. 140 Centers for Disease Control and Prevention (CDC). ( 2015 a). Guide to confirming a d iagnosis in foodborne disease. Retrieved from http://www. cdc.gov/foodsafety/outbreaks/investigatingoutbreaks/confirming_diagnosis . html . Accessed on 26 June 2015. Centers for Disease Control and Prevention (CDC). ( 2015 b ) . Surveillance for foodborne disease o utbreaks, United States, 2013, annual r eport. Atlan ta, Georgia: US Department of Health and Human Services. Center for Food Safety and Applied Nutr ition (CFSAN). (2001). Juice - g uidance for industry: the juice HACCP regulation - questions & a nswers. Retri e ved from http://www.fda.gov/Food/GuidanceRegulation/GuidanceDocumentsRegulatoryInformation/ Juice/ucm072602.htm . Accessed on 26 June 2015. Center for Food Safety and Applied Nutrit ion (CFSAN). ( 2006). Produce & plant products - commodity specific food safety guidelines for the l ettuce an d leafy greens supply chain - 1st e dition. Retri e ved from http://www.fda.gov/downloads/Food/GuidanceRegulation/ UCM169008.pdf . A ccessed on 26 June 2015. Center for Food Safety and Applied Nutrit ion (CFSAN). (200 9). Produce & plant products - dr aft guidance for industry: g uide to minimize microbial food safety h azards o f leafy g reens . Retrieved from http://www.fda.gov/Food/GuidanceRegulation/ GuidanceDocumentsRegulatoryInformation/ProducePlantProducts/ucm174200.htm . A ccessed on 26 June 2015. Center for Food Safety and Applied N utrition (CFSAN). (2014). Safe practices for food processes - chapter V. Methods to reduce/eliminate pathogens from produce and fresh - cut produce. Retrieved from http://www.fda.gov/Food/FoodScienceResearch/ SafePracticesforFoodProcesses / ucm091363.htm . Accessed on 26 June 2015. Center f or Food Safety and Applied N utrition (CFSAN). (2015). Safe practices for food processes - chapter IV. Outbreaks associated with fresh and fresh - cut produce. Incidence, growth, and survival of pathogens in f resh an d fresh - cut p roduce . Retrieved from http://www.fda.gov/Food/FoodScienceResearch/SafePracticesforFoodProcesses/ucm09126 5.htm . Accessed on 26 June 2015. Cheong, S., Lee, C., Song, S. W., Choi, W. C., Lee, C. H., & Kim, S. - J. (2009). Enteric viruses in raw vegetables and groundwater used for irrigation in South Korea. Applied and Environmental Microbiology , 75 (24), 7745 51. Chi, J., VanLeeuwen, J. A., Weersink, A., & Keefe, G. P. (2002). Direct production losses and treatment costs from bovine viral diarrhoea virus, bovine leukosis virus, Mycobacterium avium subspecies paratuberculosis, and Neospora caninum. Preventive Veterinary Medicine , 55 (2), 137 153. 141 Choi, C., Song, I., S tine, S., Pimentel, J., & Gerba, C. (2004). Role of irrigation and wastewater reuse: comparison of subsurface irrigation and furrow irrigation. Water Science and , 50 (2), 61 8. Clemens, R. (2004). The expanding U . S . market for fresh p roduce. Iowa Ag Review , 10 (1) , 1 - 4 . Colomer - Lluch, M., Jofre, J., & Muniesa, M. (2011). Antibiotic resistance genes in the bacteriophage DNA fraction of environmental samples. PloS One , 6 (3), e17549. Cook, N. ( 2013 ) . Viruses in food and water: risks, surveillance and c ontrol , 1st Ed., Woodhead Publishing Limited, Cambridge, UK. Croci, L., De Medici, D., Scalfaro, C., Fiore, A., & Toti, L. (2002). The survival of hepatitis A virus in fresh produce. International Journal of Food Microbiology , 73 (1), 29 34. Croci, L., Dubois, E., Cook, N., Medici, D ., Schultz, A. C., China, B., Rutjes, S . A., Hoorfar, J., & Van der Poel, W. H. M. (2008). Curre nt methods for extraction and concentration of enteric viruses from fresh fruit and vegetables: towards international s tandards. Food Analytical Methods , 1 (2), 73 84. Crutchfield, S. R., & Roberts, T. (2000). Food safety efforts a Food Review , 23 (202), 44 49. Crusoe, M., Edvenson, G., Fish, J., Howe, A., McDonald, E., Nahum, J., Nanlohy, K., Ortiz - Zuazaga, H., Pell, J., Simpson, J., Scott, C., Sr inivasan, R. R., Zhang, Q., & Brown, C. T. (2014). The khmer software package: e nabling efficient sequence analysis. Retrieved from http://figshare.com/articles/The_khmer_software_package_enabling _efficient_sequence_ analysis/979190 . Accessed on 3 August 2015 . C utadapt. ( 2015 ). Cutadapt - A tool that removes adapter sequences from DNA sequencing reads. Retrieved from https://code.google.com/p/cutadapt/. Accessed 3August , 2015. microbe that is antagonistic to the dysentery bacillus. Comptes mie des Sciences Paris . Davidson, G. R., Buchholz, A. L., & Ryser, E. T. (2013 ). Efficacy of commercial produce s an itizers against n ontoxigenic Escherichia coli O157:H7 during processing of iceberg lettuce in a pilot - scale leafy green processing l ine. Jour nal of Food Protection , 76 (11), 1838 1845. Davis, B. M., Moyer, K. E., Boyd, E. F., & Waldor, M. K. (2000). CTX prophages in classical biotype Vibrio cholerae: functional phage genes but dysfunctional phage genomes. Journal of Bacteriology , 182 (24), 6992 8. 142 Dawson, D. J., Paish, A. , Staffell, L. M., Seymour, I. J., & Appleton, H. (2005). Survival of viruses on fresh produce, using MS2 as a surrogate for norovirus. Journal of Applied Microbiology , 98 (1), 203 9. Deboosere, N., Pinon, A., Caudrelier, Y., Lout reul , J., Merle, G., Delobel, A., Merle, G., Perelle, S., Temmam, S., Loutreul, J., Mornin, T., Estienney, M., Belliot, G., Pothier, P., Gantzer, C., Vialette, M. (2012). Adhesion of human pathogenic enteric viruses and surrogate viruses to inert and veget al food surfaces. Food Microbiology , 32 (1), 48 - 56. Dicaprio, E., Ma, Y., Purgianto, A., Hughes, J., & Li, J. (2012). Internalization and dissemination of human norovirus and animal caliciviruses in hydroponically grown romaine lettuce. Applied and Envi ronmental Microbiology , 78 (17), 6143 52. Doultree, J. C., Druce, J. D., Birch, C. J., Bowden, D. S., & Marshall, J. A. (1999). Inactivation of feline calicivirus, a Norwalk virus surrogate. Journal of Hospital Infection , 41 (1), 51 57. Dubois, E., Hennech art, C., Deboosère, N., Merle , G., Legeay, O., Burger, C., Le Cal vé, M., Lombard, B., Ferré, V., & Traoré, O. (2006). Intra - laboratory validation of a concentration method adapted for the enumeration of infectious F - specific RNA coliphage, enterovirus, an d hepatitis A virus from inoculated leaves of salad vegetables. International Journal of Food Microbiology , 108 (2), 164 71. Dubois, E., Hennechart, C., Merle, G., Burge r, C., Hmila, N., Ruelle, S., Perelle, S., & Ferré, V. (2007). Detection and quantifica tion by real - time RT - PCR of hepatitis A virus from inoculated tap waters, salad vegetables, and soft fruits: characterization of the method performances. International Journal of Food Microbiology , 117 (2), 141 9. Ercolini, D. (2013). High - throughput seque ncing and metagenomics: moving forward in the culture - independent analysis of food microbial ecology. Applied and Environmental Microbiology , 79 (10), 3148 55. Food and Agriculture Organization of the United Nations/World Health Organizat ion (FAO/WHO). ( 20 08 ) . Viruses in food: scientific advice to support risk management activities, meeting report. Microbiological Risk Assessment Series No. 14. Rome. 151pp. Fletcher, M., Levy, M. E., & Griffin, D. D. (2000). Foodborne outbreak of group A rotavirus gastroe nteritis among college students - District of Columbia, March - April 2000. Morbidity and Mortality Weekly Report , 49 (50), 1131 1133. Fong, T. - T., & Lipp, E. K. (2005). Enteric viruses of humans and animals in aquatic environments: health risks, detection, a nd potential water quality assessment tools. Microbiology and Molecular Biology Reviews , 69 (2), 357 71. 143 Fraisse, A., Temmam, S., Deboosere, N., Guillier , L., Delobel, A., Maris, P., Vialette, M., Morin, T., & Perelle, S. (2011). Comparison of chlorine and peroxyacetic - based disinfectant to inactivate Feline calicivirus, Murine norovirus and Hepatitis A virus on lettuce. International Journal of Food Microbiology , 151 (1), 98 104. Gallagh er, S. R., & Wiley, E. A . (2008). Current protocols essential laborato ry t echniques . Hoboken, NJ, USA: John Wiley & Sons, Inc. Ganesh, B., Bányai, K., Martella, V., Jakab, F., Masachessi, G., & Kobayashi, N. (2012). Picobirnavirus infections: viral persistence and zoonotic potential. Reviews in Medical Virology , 22 (4), 245 56. Geering, A. (2007). Caulimoviridae (Plant Pararetroviruses). eLS . Gelting, R. J., & Baloch, M. A. (2012). The food - water nexus: irrigation water quality, risks to food safety, and the need for a systems - based preventive approach. Journal of Environmen tal Health , 75 (3), 40 1. Gibson, K. E., & Schwab, K. J. (2011). Tangential - flow ultrafiltration with integrated inhibition detection for recovery of surrogates and human pathogens from large - volume source water and finished drinking water. Applied and Envi ronmental Microbiology , 77 (1), 385 91. Gil, M. I., Selma, M. V, López - Gálvez, F., & Allende, A. (2009). Fresh - cut product sanitation and wash water disinfection: problems and solutions. International Journal of Food Microbiology , 134 (1 - 2), 37 45. Goldsmi th, C. S., & Miller, S. E. (2009). Modern uses of electron microscopy for detection of viruses. Clinical Microbiology Reviews , 22 (4), 552 63. Gould, L. H., Walsh, K. A , Vieira, A. R., Herman, K., Williams, I. T., Hall, A. J., & Cole, D. (2013). Surveillanc e for foodborne disease outbreaks - United States, 1998 - 2008. Morbidity , 62 (2), 1 34. Greening, G. E., Hewitt, J., & Lewis, G. D. (2002). Evaluation of integrated cell culture - PCR (C - PCR) for virological analysis of environmental samples. J ournal of Applied Microbiology , 93 (5), 745 750. Gulati, B. R., Allwood, P. B., Hedberg, C. W., & Goyal, S. M. (2001). Efficacy of commonly used disinfectants for the inactivation of calicivirus on strawberry, lettuce, and a food - contact surface. Journal o f Food Protection , 64 (9), 1430 4. Handelsman, J. (2004). Metagenomics: application of genomics to uncultured microorganisms. Microbiology and Molecular Biology Reviews , 68 (4), 669 85. 144 Harwood, V. J., Levine, A. D., Scott, T. M., Chivukula, V., Lukasik, J. , Farrah, S. R., & Rose, J. B. (2005). Validity of the indicator organism paradigm for pathogen reduction in reclaimed water and public health protection. Applied and Environmental Microbiology , 71 (6), 3163 70. Havelaar, A. H., van Olphen, M., & Drost, Y. C. (1993). F - specific RNA bacteriophages are adequate model organisms for enteric viruses in fresh water. Appl. Envir. Microbiol. , 59 (9), 2956 2962. Heaton, J. C., & Jones, K. (2008). Microbial contamination of fruit and vegetables and the behaviour of en teropathogens in the phyllosphere: a review. Journal of Applied Microbiology , 104 (3), 613 26. Hennechart, C., Merle, G., & Dubois, E. (2002). Modified concentration method for the detection of enteric viruses on fruits and vegetables by reverse transcript ase - polymerase chain reaction or cell c ulture, 65 (12), 1962 1969. Huson, D. H., Auch, A. F., Qi, J., & Schuster, S. C. (2007). MEGAN analysis of metagenomic data. Genome Research , 17 (3), 377 86. Hutin, Y., Pool, V., Cramer, E., Nainan, O. V ., Weth, J., W illiams, I. T., Goldstein, S. T., Gensheimer, K. F., Bell, B. P., Shapiro, C. N., Alter, M. J., & Margolis, H. (1999). A multistate, foodborne outbreak of hepatitis A. The New England Journal of Medicine , 340 (8), 595 602. Hyeon, J. - Y., Chon, J. - W., Park, C ., Lee, J. - B., Choi, I. - S., Kim, M. - S., & Seo, K. - H. (2011). Rapid detection method for hepatitis A virus from lettuce by a combination of filtration and integrated cell culture - real - time reverse transcription PCR. Journal of Food Protection , 74 (10), 1756 61. Illumina, Inc. ( 2015 ) . Next - generation sequencing (NGS). Retrieved from http://www.illumina.com/technology/next - generation - sequencing.html. Accessed on 26 March 2012 International Food Information Foundation Council (IFIC Foundation). ( 2014 ) . More Amer icans choosing foods, beverages based on healthfulness. Retrieved from http://www.foodinsigh t.org/articles/more - americans - c hoosing - foods - beverages - based - healthfulness. Accessed on 26 March 2015. Isakbaeva, E. T., Widdowson, M. - A., Beard, R. S., Bulens, S. N. , Mullins, J., Monroe, S. S., Bresee, J., Sassano, P., Cramer, E. H., & Glass, R. I. (2005). Norovirus transmission on cruise ship. Emerging Infectious Diseases , 11 (1), 154 8. Johnson, P. T. (1984). Viral diseases of marine invertebrates. Helgoländer M eeresuntersuchungen , 37 (1 - 4), 65 98. 145 Kim, M. - S., Park, E. - J., Roh, S. W., & Ba e, J. - W. (2011). Diversity and abundance of single - stranded DNA viruses in human f eces. Applied and Environmental Microbiology , 77 (22), 8062 8070. Kircher, M., & Kelso, J. (201 0). Hi gh - throughput DNA sequencing - c oncepts and limitations. BioEssays , 32 (6), 524 536. Koopmans, M., & Duizer, E. (2004). Foodborne viruses: an emerging problem. International Journal of Food Microbiology , 90 (1), 23 41. Kotula, K. L., Kotula, A. W., R ose, B. E., Pierson, C. J., & Camp, M. (1997). Reduction of aqueous chlorine by organic m aterial. Journal of Food Protection , 60 (3), 276 282 . Kroneman, A. , Verhoef, L., Harris, J., Vennema, H., D uizer, E., van Duynhoven, Y., Gray, J., Iturriza, M., Böttige r, B., Falkenhorst, G., Johnsen, C., von Bonsdorff, C. H., Maunula, L., Kuusi, M., Pothier, P., Gallay, A., Schreier, E., Hö hne, M., Koch, J., Szücs, G., Reuter, G., Krisztalovics, K., Lynch, M., McKeown, P., Foley, B., Coughlan, S., Ruggeri, F. M., Di Ba rtolo, I., Vainio, K., Isakbaeva, E., Poljsak - Prijatelj, M., Hocevar Grom, A., Zimsek Mijovski, J., Bosch, A., Buesa, J., Sanchez Fauquier, A., Hernandéz - Pezzi, G., Hedlund, K. - O., & Koopmans, M. (2008). Analysis of integrated virological and epidemiologic al reports of norovirus outbreaks collected within the foodborne v iruses in Europe network from 1 July 2001 to 30 June 2006. Journal of Clinical Microbiology , 46 (9), 2959 65. Kropinski, A. M., Mazzocco, A., Waddell, T. E., Lingohr, E., & Johnson, R. P. (2 009). Enumeration of bacteriophages by double agar overlay plaque assay. Bacteriophages, M ethods in Molecular Biology (Clifton, N.J.) , ( pp. 69 76 ) . Humana Press. Kunin, V., Copeland, A., Lapidus, A., Mavromatis, K., & Hugenholtz, P. (2008). A bioinformati Microbiology and Molecular Biology Reviews, 72 (4), 557 78. Kutter, E., & Sulakvelidze, A. (2004). Bacteriophages: biology and a pplications . CRC Press . Lacey, L. , Frutos, R., Kaya, H . , & Vail, P. (2001). Insect p athogens as biological control agents: do they have a f uture? Biological Control , 21 (3), 230 248. Leclerc, H., Edberg, S., Pierzo, V., & Delattre, J. M. (2000). Bacteriophages as indicators of enteric viruses and public health risk in groundwaters. Journal of Applied Microbiology , 88 (1), 5 21. Leclerc, H., Edberg, S., Pierzo, V., & Delattre, J. M. (2001). Bacteriophages as indicators of enteric viruses and public health risk in groundwaters. Journal of Applied Microbiology , 88 (1), 5 21. 146 Lee, H. K., & Jeong, Y. S. (20 04). Comparison of total culturable virus assay and multiplex integrated cell culture - PCR for reliability of waterborne virus detection. Applied and Environmental Microbiology , 70 (6), 3632 6. Leff, J. W., & Fierer, N. (2013). Bacterial communities a ssoci a ted with the surfaces of fresh fruits and v egetables. PLoS ONE , 8 (3), 1 9. Leggitt, P. R., & J aykus, L. - A. (2000). Detection methods for human enteric viruses in representative f oods. Journal of Food Protection , 63 (12), 1738 44. Leland, D. S., & Ginocchio , C. C. (2007). Role of cell culture for virus detection in the age of technology. Clinical Microbiology Reviews , 20 (1), 49 78. Lindgreen, S. (2012). Adapter R emoval: easy cleaning of n ext - generation sequencing r eads. BMC Research Notes , 5 (1), 337. Liu, L ., Li, Y., Li, S., Hu, N., He, Y., Pong, R., Lin, D., Lu, L., & Law, M. (2012). Comparison of next - generation sequencing systems. Journal of Biomedicine and Biotechnology , 2012 . Liu, P., Hill, V. R., Hahn, D., Johnson, T. B., Pan, Y., Jothikumar, N., & Mo e, C. L. (2012). Hollow - fiber ultrafiltration for simultaneous recovery of viruses, bacteria and parasites from reclaimed water. Journal of Microbiological Methods , 88 (1), 155 61 . Liu, S., Vijayendran, D., & Bonning, B. C. (2011). Next generation sequencin g technologies for insect virus discovery. Viruses , 3 (10), 1849 1869. Lopez - Velasco, G., Welbaum, G. E., Boyer, R. R., Mane, S. P., & Ponder, M. A . (2011). Changes in spinach phylloepiphytic bacteria communities following minimal processing and refrigerat ed storage described using pyrosequencing of 16S rRNA amplicons. Journal of Applied Microbiology , 110 (5), 1203 1214. Love, D. C., Casteel, M. J., Meschke, J. S., & Sobsey, M. D. (2008). Methods for recovery of hepatitis A virus (HAV) and other viruses fr om processed foods and detection of HAV by nested RT - PCR and TaqMan RT - PCR. International Journal of Food Microbiology , 126 (1 - 2), 221 6. Luscombe, N. M., Greenbaum, D., & Gerstein, M. (2001). What is bioinformatics? A proposed definition and overview of t he field. Methods of Information in Medicine , 40 (4), 346 58. Lynch, M. F., Tauxe, R. V, & Hedberg, C. W. (2009). The growing burden of foodborne outbreaks due to contaminated fresh produce: risks and opportunities. Epidemiology and Infection , 137 (3), 307 1 5. Mehle, N., & Ravnikar, M. (2012). Plant viruses in aqueous environment - Survival, water mediated transmission and detection. Water Research , 46 (16), 4902 4917. 147 Meng, X. - J. (2011). From barnyard to food table: the omnipresence of hepatitis E virus an d risk for zoonotic infection and food safety. Virus Research , 161 (1), 23 30. Mokili, J. L., Rohwer, F., & Dutilh, B. E. (2012). Metagenomics and future perspectives in virus discovery. Current Opinion in Virology , 2 (1), 63 77. Moore, N. E., Wang, J., He witt, J., Croucher, D., Will iamson, D. A., Paine, S., Yen, S., Greening, G. E., & Hall, J. (2015). Metagenomic analysis of viruses in feces from unsolved outbreaks of gastroenteritis in h umans. Journal of Clinical Microbiology , 53 (1), 15 21. National Food Processors Association (NFPA). (2001). International Fresh - cut Produce Association, United Fresh Fruit and Vegetable Association. Field Cored Lettuce Best Practices, NFPA, Washington, DC. Newell, D. G., Koopmans, M., Verhoef, L., Duizer, E., Aidara - Kane , A., Sprong, H., Opsteeh, M., Langelaar, M., Threfall, J., Scheutz, F., van der Giessen, J., & Kruse, H. (2010). Food - borne diseases - the challenges of 20 years ago still persist while new ones continue to emerge. International Journal of Food Microbiolo gy , 139, 3 15. Food - Borne Viruses: Progress and Challenges (pp. 87 115). American Soceity of Microbiology , 1752 N Street NW, Washington, DC USA. Oliver, S. P., Pa tel, D. A , Callaway, T. R., & Torrence, M. E. (2009). ASAS centennial paper: d evelopments and future outlook for preharvest food safety. Journal of Animal Science , 87 (1), 419 37. Ottesen, A. R., González Peña, A., White, J. R., Pettengil l, J. B., Li, C., Allard, S., Rideout, S., Allard, M., Hill, T., Evans, P., Strain, E., Musser, S., Knight, R., & Brown, E. (2013). Baseline survey of the anatomical microbial ecology of an important food plant: Solanum lycopersicum (tomato). BMC Microbiology , 13 (1), 114. P ainter, J. A ., Hoekstra, R. M., Ayers, T., Tauxe, R. V., Braden, C. R., Angulo, F. J., & Griffin, P. M. (2013). Attribution of foodborne illnesses, hospitalizations, and deaths to food commodities by using outbreak data, United States, 1998 - 2008. Emerging Infectious Diseases , 19 (3), 407 415. Park, E. - J., Kim, K. - H., Abell, G. C. J., Kim, M. - S., Roh, S. W., & Bae, J. - W. (2011). Metagenomic analysis of the viral communities in fermented foods. Applied and Environmental Microbiology , 77 (4), 1284 91. Peng, Y. , Leung, H. C. M., Yiu, S. M., & Chin, F. Y. L. (2012). IDBA - UD: A de novo assembler for single - cell and metagenomic sequencing data with highly uneven depth. Bioinformatics , 28 (11), 1420 1428. 148 Pimentel, D., McNair, S., Janecka, J., Wightman, J., Simmond ., Wong, E., Russel, L., Zern, J., Aquino, T., & Tsomondo, T. (2001). Economic and environmental threats of alien plant, animal, and microbe invasions. Agriculture, Ecosystems & Environment , 84 (1), 1 20. P ollack, S. L. (2001). Consumer demand for fruit and v egetables : the U.S. example. Changing St ructure of Global Food and Trade WRS ( 49 54). Reynolds, K. A. (2004). Integrated cell culture/PCR for detection of enteric viruses in environmental samples. Methods in Molecular Biology (Clifton , N.J.) . (pp. 69 78 ) . Humana Press. Roche Diagnostics Corporation. ( 2015 ) . GS FLX+ System. Retrieved from http://454.com/products/gs - FLX - system/index.asp . Accessed on 26 March 2015. Rodríguez - Lázaro, D., Cook, N., Ruggeri, F. M., Sellwood, J., Nass er, A., Nascimento, M. S. J., Saiz, J. C., Rzezutka, A., Bosch, A., Gironés, R., Carducci, A., - Valcarce, M., Vantarakis, A., von Bonsdorff, C. - H., de Roda Husman, A. M., Hernández, M., & van der Poel, W. H . M. (2012). Virus hazards from food, water and other contaminated environments. FEMS Microbiology Reviews , 36 (4), 786 814. Rosario, K., & Breitbart, M. (2011). Exploring the viral world through metagenomics. Current Opinion in Virology , 1 (4), 289 297. R osario, K., Nilsson, C., Lim, Y. W., Ruan, Y., & Breitbart, M. (2009). Metagenomic analysis of viruses in reclaimed water. Environmental Microbiology , 11 (11), 2806 20. Ross, E. M., Moate, P. J., Bath, C. R., Davidson, S. E., Sawbri dge, T. I., Guthridge, K . M., Cocks, B. G., & Hayes, B. J. (2012). High throughput whole rumen metagenome profiling using untargeted massively parallel sequencing. BMC Genetics , 13 (1), 53. Ross, E. M., Petrovski, S., Moate, P. J., & Hayes, B. J. (2013). Metagenomics of rumen bac teriophage from thirteen lactating dairy cattle. BMC Microbiology , 13 (1), 242. Rybicki, E. P. (2015). A top t en list for economically important plant viruses. Archives of Virology , 160 (1), 17 20. Rzezutka, A., & Cook, N. (2004). Survival of human enteric viruses in the environment and food. FEMS Microbiology Reviews , 28 (4), 441 53. Sánchez, G., Elizaquível, P., & Aznar, R. (2012). A single method for recovery and concentration of enteric viruses and bacteria from fresh - cut vegetables. International Journ al of Food Microbiology , 152 (1 - 2), 9 13. Scharff, R. L. (2012). Economic burden from health losses due to foodborne illness in the United States. Journal of Food Protection , 75 (1), 123 31. 149 Seo, K., & Frank, J. (1999). Attachment of Escherichia coli bacterial viability in response to chlorine treatment as demonstrated by using confocal scanning laser microscopy. J ournal of Food Protection , 62 (1), 3 9. Seymour, I. J., & Appleton, H. (2001). Foodborne viruses and f resh produce. Journal of Applied Microbiology , 91 (5), 759 73. Shang, C., & Blatchley, E. R. (2001). Chlorination of pure bacterial cultures in aqueous solution. Water Research , 35 (1), 244 254. Sharma, M. (2013). Lytic bacteriophages: Potential interventio ns against enteric bacterial pathogens on produce. Bacteriophage , 3 (2), e25518. Sivapalasingam, S., Friedman, C. R., Cohen, L., & Tauxe, R. V. (2004). Fresh produce: a growing cause of outbreaks of foodborne illness in the United States, 1973 through 1997 . Journal of Food Protection , 67 (10), 2342 53. Smith, C. M., & Hill, V. R. (2009). Dead - end hollow - fiber ultrafiltration for recovery of diverse microbes from water. Applied and Environmental Microbiology , 75 (16), 5284 9. Song, I., Stine, S. W., Choi, C. Y., & Gerb a, C. P. (2006). Comparison of crop contamination by microorganisms during subsurface drip and furrow i rrigation. Journal of Environmental Engineering , 132 (10), 1243 1248. Steele, M., & Odumeru, J. (2004). Irrigation water as source of foodborn e pathogens on fruit and vegetables. Journal of Food Protection , 67 (12), 2839 2849. Stine, S. W., Song, I., Choi, C. Y., & Gerba, C. P. (2005). Application of microbial risk assessment to the development of standards for enteric pathogens in water used to irrigate fresh produce. Journal of Food Protection , 68 (5), 913 918. Taormina, P. J., Beuchat, L. R., Erickson, M. C., Ma, L., Zhang, G., & Doyle, M. P. (2009). Transfer of Escherichia coli O157:H7 to iceberg lettuce via simulated field c oring. Journal of F ood Protection , 72 (3), 456 684. Telias, A., White, J. R., Pahl, D. M., Ottesen, A. R., & Walsh, C. S. (2011). Bacterial community diversity and variation in spray water sources and the tomato fruit surface. BMC Microbiology , 11 (1), 81. United States Depar tment of Agriculture (USDA). ( 2014 ). Economic research service: food availability (p er capita) data system: s umm ary f indings. United States Department of Agriculture website . Retrieved from http://www.ers.usda.gov/data - products/food - availability - (per - capita) - data - system/summary - findings.aspx . Accessed on 26 March 2015. 150 United States Environme ntal Protection Agency (EPA) (2001). Method 1602: male - specific (F+) and s omatic coliphage in w ater by single agar layer p rocedure. EPA 821 - R - 98 - 029. Office of Water, Washington, DC. United S tates Environmental Protection Agency (EPA) (2012). Guidelines for Water Reuse. EPA - 600 - R - 12 - 618. Office of Water, Washingtion DC. Varma, A., & Malathi , V. G. (2003). Emerging geminivirus problems: a serious threat to crop production. Annals of Applied B iology , 142 (2), 145 164. Vázquez - Castellanos, J. F., García - López, R., Pérez - Brocal, V., Pignatelli, M., & Moya, A. (2014). Comparison of different assembly and annotation tools on analysis of simulated viral metagenomic communities in the gut. BMC Genomi cs , 15 (1), 37. Vega, E., Garland, J., & Pillai, S. D. (2008). Electrostatic forces control nonspecific virus attachment to lettuce. Journal of Food Protection , 71 (3), 522 9. Vega, E., Smith, J., Garland, J., Matos, A., & Pillaii, S. D. (2005). Variabilit y of virus attachment patterns to butterhead lettuce. Journal of Food Protection , 68 (10), 2112 7. Voller, A., Bartlett, A., Bidwell, D. E., Clark, M. F., & Adams, A. N. (1976). The detection of viruses by enzyme - linked immunosorbent a ssay (ELISA). Journal of General Virology , 33 (1), 165 167. Wang, D., Urisman, A., Liu, Y. T., Springer, M., Ksiazek, T. G., Erdman, D. D., Mardis, E. R., Hickenbotham, M., Magrini, V., Eldred, J., Latreille, J. P., Wilson, R. K., Ganem, D., & DeRisi, J. L. (2003). Viral discove ry and sequence recovery using DNA microarrays. PLoS Biology , 1 (2). Ward, B. K., & Irving, L. G. (1987). Virus survival on vegetables spray - irrigated with wastewater. Water Research , 21 (1), 57 63. Wei, J., Jin, Y., Sims, T., & Kniel, K. E. (2010). Manure - and biosolids - resident murine norovirus 1 attachment to and internalization by Romaine lettuce. Applied and Environmental Microbiology , 76 (2), 578 83. Wheeler, C., Vogt, T. M., Armstrong, G. L., Vaughan, G., Weltman , A., Nainan, O. V., Dato, V., Xia, G. , Waller, K., Amon, J., Lee, T. M., Highbaugh - Battle, A., Hembree, C., Evenson, S., Ruta, M. A., Williams, I. T., Fiore, A. E., & Bell, B. P. (2005). An outbreak of hepatitis A associated with green o nions. The New England Journal of Medicine , 353 (9), 890 897. Widdowson, M. - A., Sulka, A., Bulens, S. N., Beard, R. S. , Chaves, S. S., Hammond, R., Salehi, E. D. P., Swanson, E., Totaro, J., Woron, R., Mead, P. S., Bresee, J. S., Monroe, S. S., & Glass, R. I. (2005). Norovirus and foodborne disease, United State s, 1991 - 2000. Emerging Infectious Diseases , 11 (1), 95 102. 151 Wilson, A. E., Sarnelle, O., & Tillmanns, A. R. (2006). Effects of cyanobacterial toxicity and morphology on the population growth of freshwater zooplankton: Meta - analyses of laboratory experiment s. Limnology and Oceanography , 51 (4), 1915 1924. Wylie, K. M., Weinstock, G. M., & Storch, G. A . (2013). Virome genomics: A tool for defining the human virome. Current Opinion in Microbiology , 16 (4), 479 484. Yépiz - Gómez, M. S., Gerba, C. P., & Bri ght, K. R. (2013). Survival of respiratory viruses on fresh pr oduce. Food and Environmental Virology , 5 (3), 150 156.