Project Summary

High-Resolution Serotype Population Analyses in Pre-Harvest Cattle to Reveal Salmonella Transmission Networks 

Principle Investigator(s):
Bryan Larsen1, Christian Gunning2, Diego Casas3, David Vargas3, Tayler Hayes3, Tyson Brown4, Angie Siemens4, Dale Woerner3, Nikki Shariat1
1Department of Population Health, University of Georgia
2Odum School of Ecology, University of Georgia
3Department of Animal and Food Sciences, Texas Tech University
4Cargill Inc.
Completion Date:
September 2022

  • CRISPR-SeroSeq analysis of 609 samples collected in commercial feedlots showed that nearly 80% of samples contain two or more serotypes, with an average of 2.9 serotypes per sample (range 1-10 serotypes per sample). This demonstrates that traditional methods for monitoring Salmonella serotypes are not sufficient to assess Salmonella serotype dynamics in beef production.
  • 20/23 serotypes occurred more frequently as ‘minority serotypes’ where they were outnumbered by another serotype in the same sample. Minority serotypes are not typically detected by traditional culture.
  • Salmonella prevalence, overall number of Salmonella serotypes, and the number of serotypes per sample was highest in the summer, in one feedlot, and in the upper surface layers of feedlot pens. Shipping and receiving pens harbored highly complex Salmonella serotype populations, likely because of the high levels of cattle transfer through these pens.
  • Boot sock samples of the manure pad were most likely to be Salmonella positive and the best sample type to capture the whole Salmonella serotype diversity within a pen.


Salmonella is a leading bacterial foodborne pathogen in the United States, and 6.2% of human cases are linked to consumption of contaminated beef products. Salmonella transmission within feedlot ecosystems is complex, with Salmonella moving bidirectionally between cattle and their environments. Multiple factors are attributed to influencing Salmonella prevalence, including the feedlot environment, cattle breed and source, animal health, seasonality, and climate. Mounting evidence exists that shows Salmonella prevalence differs among feedlots, even when the feedlots are within the same geographical/climatic region. Data collected over several years by a processor shows that when suppliers are rank ordered according to subiliac peripheral lymph node Salmonella prevalence, feedlots can be categorized as “high” or “low” year-over-year. This illustrates that feedlot environment or production practices play a key role in Salmonella dynamics and an understanding of these dynamics is required to improve food safety in this important sector.   

Salmonella serotypes are defined by the identities of the cell surface O and H antigens. Many serotypes exhibit differences in important phenotypes, such as their association with human illness, or their host restriction/adaptation. These differences suggest that different serotypes are better adapted to different niches, and this confounds our ability to precisely understand Salmonella transmission dynamics. Conventional culture-based Salmonella surveillance culminates in picking and serotyping a single colony off an agar plate, and concluding that this serotype constitutes the entire population of Salmonella in a particular sample. This is biased toward only identifying the most abundant serotype in a population (or one that grows best under those media conditions) while serotypes that are less abundant remain hidden and undetected (i.e. “minority” serotypes). Where clinically important serotypes are undetected, traditional surveillance underestimates the presence of Salmonella serotypes of the greatest food safety concern. Deep serotyping of Salmonella by CRISPR-SeroSeq overcomes this limitation by resolving all serotypes present in a single sample and also determines the relative frequency of each serotype. This method has revealed serotypes at orders of magnitude greater than logistically possible by picking colonies off a plate and has shown that multi-serotype populations of Salmonella occur frequently in poultry, freshwater environments, and cattle in an experimental feedlot. Given phenotypic and risk differences among different serotypes, a need exists to resolve all serotypes in mixed populations to understand which serotypes better survive in different niches and how they are transmitted.

Over one year, four feedlots were visited four times and at each visit, eight pens were sampled (total, n=128 pens) for various surface, soil core, and other environments (total samples, n=896). Shipping and receiving pens were also swabbed at each visit. Two feedlots were selected and two pens in each feedlot were sampled following placement of cattle in the winter. This was then repeated in the same two feedlots but different pens for cattle placed in the summer. Six pairs of samples were collected in each pen: boot socks to sample the dried manure pad, and fresh feces. Sampling was performed at time of placement (day 0), then days 15, 30, 45, 60, 90, and 120. Salmonella prevalence for both objectives was determined by culture and confirmed by qPCR. Salmonella-positive samples were profiled by CRISPR-SeroSeq and the relative frequency of each serotype was recorded and presented in heatmap form. Several statistical analyses were used, including logistic regression, principal coordinates analysis, Bray-Curtis and Jaccard population analyses, odds ratio and Fisher test, paired Wilcoxon test. All analyses were done in R.

Salmonella prevalence and sample complexity (number of serotypes per sample) were highest in a single feedlot across all four seasons. This feedlot was also the largest of the four feedlots surveyed. Overall, Salmonella prevalence was highest in summer and fall and for cattle placed in the summer compared to the winter. 80% of samples contained two or more serotypes, with one sample containing 10 serotypes. The seven serotypes found that are also in the CDC Top 10 accounted for one third (34.3%) of all serotype detections. It is estimated this would be 10% using traditional microbiology techniques and colony picking. Samples collected from upper levels (manure pad, fecal, soil surface) within a pen were more likely to contain Salmonella and more likely to contain more serotypes than cognate samples collected from the same pen but lower in the soil. The presence of Salmonella serotypes in lower soil layers suggests that these serotypes can persist deep in the soil. Shipping and receiving pens harbored large numbers of different Salmonella serotypes. The longitudinal study showed that serotypes present at placement tended to persist in the pen throughout finishing.

Salmonella remains a concern for the beef industry. Current federal performance standards are being renewed for poultry and pork and will likely focus on beef products in the near future. An improved understanding of transmission dynamics and serotype persistence at pre-harvest can guide appropriate Salmonella interventions and management decisions during finishing that will reduce the Salmonella burden at processing, ultimately improving beef safety. This study highlights that Salmonella dynamics in beef feedlots are remarkably complex and that when present, occurs four out of every five times in complex populations of multiple serotypes. This is in contrast to vertically integrated food animal production systems (i.e. poultry) where serotype diversity is not as rich on a per unit basis, likely because animals in a single feedlot pen can originate from multiple different sources, each with different potential Salmonella exposures. Further, because feedlots are open systems, Salmonella can be transmitted into the environment by weather events (e.g., rain, wind) or by animals (e.g., flies, birds), and this may add to the diversity of Salmonella observed in these environments. Of the 37 serotypes found, most were not commonly associated with human illness. Seven were in the CDC Top 10, and these accounted for 1/3rd of all serotype detections. To be able to use appropriate interventions, it is important for processors to know what is in their production systems and deep serotyping by CRISPR-SeroSeq provides a more accurate assessment of Salmonella.