Feral pigs are also attracted to agricultural areas for food, water and mates, which facilitates direct or indirect pathogen transmission. States at greatest risk for increased contact between feral and domestic pigs and consequential potential disease transmission, are those regions containing large populations of feral pigs living near outdoor-raised domestic pigs reared with low levels of bio-security. Despite multiple control and eradication efforts, California has one of the largest and widest distributions of feral pig populations, which continues to expand. High-risk contact areas between feral pigs and outdoor-based swine could have an important role in the spread of future emerging or reemerging diseases, including TBD, which could negatively impact California, the top agricultural production state in the US. The risk map built in Chapter 2 determined that more than 40% of the 305 identified OPO in California are situated within suitable feral pig habitat areas. Both feral and domestic swine are reservoirs for STEC. A study identified feral pigs in California as reservoirs of E. coli O157:H7, with prevalence ranging from 5.0% to 23.4%, depending upon sample type. A 2018 Georgia study, another state in the US with large feral pig populations, black plastic nursery pots reported 19.5% STEC prevalence in feral pigs. In Chapter 1, Patterson et al reported a STEC prevalence of 5.59% in pigs reared outdoors on small-scale diversified farms in California.

Moreover, serogroups O26 and O103 were found in positive STEC samples in Chapter 1, which are listed in the top seven serogroups that account for over 95% of human STEC illness in the US. Foodborne STEC infections have previously occurred in humans ingesting contaminated pig products. For instance, outbreaks of E. coli O157:H7 occurred in Canada in 2011 and 2014 from consumption of contaminated pork products. Although pork is not currently considered a major source of food borne infection in the US, many studies recognize the importance of STEC maintained in both domestic pigs and feral swine in the US and internationally. The objectives of this study focused on a) measuring the prevalence of STEC in outdoor raised domestic pigs and feral pigs located near these OPO, and b) analyzing risk factors for the presence of STEC on OPO that operate near feral pig populations in six high-risk California counties, as determined by the risk map built in Chapter 2. We conducted a cross-sectional study between February and August 2018 to collect fecal samples from outdoor-raised domestic pigs and feral pigs in six high-risk California counties. Sample collection was targeted to six high-risk counties that had a higher likelihood of feral pig to outdoor-raised pig contact, based on the risk map built in Chapter 2 and communication with landowners that had feral pig presence on their land.

OPO enrollment criteria for this study included 1) reared domestic pigs outdoors in one of the six targeted counties; 2) willingness to participate; and 3) farm owners had seen evidence of feral pigs on their property, or their farm was located near suitable feral pig habitat, according to the risk map built in Chapter 2. Recruitment techniques included personal farm visits, previous working or research connections, farmers markets, and agricultural festivals. Once OPO were enrolled, we identified feral pig locations on or near those farms or at least within the same county. Feral pig locations were identified through conversations with landowners, hunters, University of California Cooperative Extension advisors or United States Department of Agriculture Wildlife Services staff, who conduct disease surveillance in feral pigs in some California counties. We collected domestic swine fecal samples from participating OPO, including ones that also had feral pig presence on their farm. To collect feral pig fecal samples, we worked with landowners to identify known locations on their properties, then we looked for signs of feral pig presence, such as wallowing areas, swine footprints and/or rooting for food. Once feral pig areas were identified, we searched for fecal samples. A feral pig fecal sample was considered authentic if the fecal pile resembled typical pig feces, was surrounded by at least of one of the feral pig presence signs listed above and there were no other livestock within the collection pasture that could contaminate a sample. All study sites were visited once, except for one private ranch that had high feral pig presence throughout the study period and was visited three times.

Based on previous studies, we assumed a STEC prevalence of 5% for outdoor-raised pigs and 10% for feral pigs with a 10% precision error for both, which resulted in needing approximately 204 domestic pig samples and 72 feral pig samples. Sample size per farm was calculated by proportional stratified methods. The total number of samples collected per farm was based on total pig count and the number inside each paddock, pen or pasture. Fresh fecal samples were collected from the ground or from a feral pig’s colon, if it had been freshly hunted. Fecal samples were gathered with gloves and placed into sterile cups with sterilized wooden tongue depressors Samples were placed into a cooler containing ice packs and brought to the lab for processing within 24-48 hrs. All participants were asked to complete a questionnaire that included topics regarding known nearest feral pig locations, farm demographics and domestic pig health, as related to each type of study participant. The survey instrument and protocols were reviewed by the Institutional Review Board of the University of California-Davis . A directed acyclic graph was built to assess key questions to include in the questionnaire. Once the study participant completed the questionnaire, they were sent a $30 gift card as a thank you gift for assisting us with the study and as a motivation to complete the survey. Participants completed the questionnaire via phone, email or mail. If missing data was identified, then a follow-up call or email was initiated to gather these answers. All fecal samples were screened for E. coli O157:H7 and non-O157 STEC and tested for stx1 and stx2 genes. Upon arrival to the laboratory, fecal samples were cultured for STEC, using a modified version of a previous protocol. Briefly, for each fecal sample, a Tryptic Soy Broth enrichment was performed for detection of non-O157 STEC and E. coli O157:H7.Ten grams of fecal material was weighed and added to a pre-refrigerated 24 oz Whirl-Pak sterile bag filled with 90 mL of TSB and manually homogenized for one minute. Then samples were incubated in a shaking incubator at 100 RPM and held at 4°C. A Multitron programmable shaking incubator was employed in this study. STEC isolates were sent for whole genome sequencing to the University of California, Davis Genome Center, after PCR confirmation, clean-up and DNA extraction. Briefly, isolates were grown overnight aerobically at 37°C in autoclaved 15 mL culture tubes containing 10 mL Brain Heart Infusion broth. DNA was then extracted according to the DNeasy Blood and Tissue Kit . To ensure adequate purity for DNA sequencing, eluted DNA was purified according to the Zymo Quick-DNA Miniprep Kit . 30 µL of purified DNA was eluted for each isolate into a sterile 2 mL micro centrifuge tube. DNA quantification was conducted for all isolates using a NanoDrop OneC . DNA samples were stored at -80°C.The raw reads were pre-processed using HTStream to remove contamination, remove duplicates, overlap reads, and trim based on quality and length. Next, the processed reads were assembled using SPAdes . BUSCO was utilized to check assembly completeness and find common single-copy orthologs across all of the isolates. Using common sequences, “pseudo-genomes” were created for each isolate, after which mafft was used to create a multiple alignment across all of them. Custom R code was used to collate the results of each of the finders into a superset of the hits for each of the analyses. Using the supersets, 30 plant pot heatmaps were generated across all the isolates and genes.Data analysis was conducted using generalized linear mixed models to identify significant risk factors for the presence of STEC in outdoor-raised pigs . Univariate analysis was used to initially assess the distribution of variables. During bivariate analysis, variables with low variability, small cell sizes , or large standard errors were either modified, collapsed if appropriate, or discarded from model building. Correlations between variables were ascertained using the Spearman’s rank correlation coefficient. Possible confounders were identified using a directed acyclic graph and then included in models to assess significant changes in the odds ratio.

The sample size in this study was too small to detect effect modifiers. Manual two-way stepwise variable selection was employed for model building, using add1 and drop1 functions in the stats R package. Models were built using the glmer function from the lme4 package, with farm as a random affect. Variance inflation factors measured multicollinearity within each model. Top models were compared, and a final model was chosen based on the lowest Akaike Information Criterion and smallest deviance. Model diagnostics were conducted on final models using the DHARMa package. Intraclass correlation was calculated. Odds ratios and 95% confidence intervals were calculated for all variables in the final model. All data analysis was performed using R Statistic Software version 1.4.1036 ©.The entire study included 17 farms or ranches, but not including the two feral pig samples collected by USDA-WS, of the remaining 16 participants: 56.25% were diversified farms ; 18.75% raised multiple types of livestock, but no crops ; 12.5% reared pigs only ; and the remaining two ranches, were private landowners who did not raise domestic pigs, but had feral pig presence on their land . Of the 16 participants, 62.50% also raised poultry, 25.00% sheep, 31.25% cattle, 31.25% goats, and 12.5% equine. Domestic swine diseases reported by farmers during the 2018 study period included pneumonia, diarrhea, PRRS and a non-diagnosed respiratory condition; each of these was reported on one farm each. Fourteen of the 16 questionnaires were completed, of which 11 were for OPO. The two incomplete questionnaires belonged to the two participants who only raised swine, not crops or other livestock. None of the answering 11 OPO had direct bordering neighbors who raised domestic swine. Five of the 11 OPO reared their domestic swine outdoors with access to wild areas and approximately half of the responding 11 OPO allowed visitors direct contact with pigs. Five of the 11 OPO had seen evidence of feral pig presence on their farm with four stating that feral pigs had direct contact with their domestic pigs in pastures, pens or barns. Of the five that had feral pig presence on their farms, three witnessed feral swine monthly and two observed them on a weekly basis. The number of feral pigs reported per farm owner ranged from a minimum of one up to 100 pigs and the maximum number observed ranged from 15 up to 300. Of the remaining six OPO with no feral pig presence on their farms, five had seen feral pigs in their counties less than five miles from their operation and only one OPO had never seen signs of feral pigs in their county. Of the six targeted counties in this study, we enrolled one to four OPO and one to three feral pig locations per county. Although we were able to enroll OPO in each of the targeted six counties, feral pig samples were unattainable in San Mateo, Nevada and Monterey counties. Feral pig samples were collected in Mendocino, Yolo, and Sonoma. This study determined STEC prevalence for feral pigs and domestic pigs raised outside near feral pig locations in high-risk California counties and assessed risk factors associated with the presence of STEC on OPO. We estimated an overall STEC prevalence in outdoor-raised pigs of 20.13%, and 11.63% in feral pigs. Also, serotypes implicated in severe human disease were identified in fecal samples through WGS. Significant risk factors associated with STEC presence on OPO included age of swine sampled , the distance to the nearest surface water, and whether a farm raised domestic swine with access to a wild area . The non-O157 STEC prevalence in this study was 18.12% in domestic pigs reared outdoors and 9.30% in feral pigs. E. coli O157:H7 results were 2.33% in feral pigs and 2.01% in domestic pigs. Prevalence of E. coli O157:H7 and non-O157 STEC in swine varies greatly worldwide, and US studies measuring STEC in OPO are sparse. Although STEC was identified in outdoor raised domestic swine in both Chapters 1 and 3, the overall prevalence was much larger in Chapter 3 versus Chapter 1 .