Data science and infectious disease ecology applications to preventing foodborne salmonellosis
USDA ARS, US Meat Animal Research Center, Meat Safety and Quality Research Unit
Advances in microbial detection and DNA sequencing technology have allowed inroads to the field of predictive microbiology. The tools presently in use, as well as those on the horizon, provide an opportunity to examine infectious diseases in ways previously not possible. While the factors that drive the emergence of a new pathogen or strain are deeply complex and interconnected, they can be succinctly summarized in three groups: host factors, pathogen factors, and environmental factors. Disease emergence can only occur when these three factors align together in space and time to create a "perfect storm." My research goal is to exploit these new technological advances along with machine learning frameworks and infectious disease ecology theory to create useful, usable, and used tools that decrease the risk of salmonellosis from meat products. I work with a diverse team of scientists as well as stakeholders in industry, the regulatory sphere, and the public to achieve these goals.
Current projects include:
1. Sources of error in Salmonella detection, quantification, and characterization in the farm-to-fork pipeline
2. Leveraging machine learning to identify and control environmental predictors of Salmonella risk in the pre-harvest sphere
3. Historical analysis of Salmonella using pre-existing data: trends in outbreaks, serotypes, genomics, and environmental factors
Check out the US Meat Animal Research Center.
Current projects include:
1. Sources of error in Salmonella detection, quantification, and characterization in the farm-to-fork pipeline
2. Leveraging machine learning to identify and control environmental predictors of Salmonella risk in the pre-harvest sphere
3. Historical analysis of Salmonella using pre-existing data: trends in outbreaks, serotypes, genomics, and environmental factors
Check out the US Meat Animal Research Center.
Postdoctoral Research: Integrated food safety systems to reduce foodborne salmonellosis
USDA ARS/SCINet: Meat Safety and Quality Research Unit
Salmonella causes over 1 million illnesses and nearly 400 deaths each year. Despite millions of dollars spent annually and a decrease in Salmonella on meat and poultry products, Salmonella illness rates have not decreased in the last two decades. Our team aims to reduce Salmonella in the meat supply chain by:
1) developing detection assays that are rapid, quantitative, determine serotype, and evaluate virulence;
2) combining data from (1) with whole genome sequencing, metagenomic data, wet lab research, and field studies to generate an understanding of Salmonella population genetics, microbial ecology, and transmission through commodities and their environments;
3) developing effective and affordable pre- and post-harvest interventions for use by commercial producers;
4) integrating data from these studies, data from monitoring programs, and machine learning to create decision support tools for each commodity, production stage, and company;
5) establishing communications and outreach efforts that allow stakeholders to provide feedback to further hone research priorities and facilitate implementation of the program.
I support the team by providing overall project management, data analytics, and data management, and serve as a project leader for Objective 4.
Check out the Salmonella Grand Challenge.
1) developing detection assays that are rapid, quantitative, determine serotype, and evaluate virulence;
2) combining data from (1) with whole genome sequencing, metagenomic data, wet lab research, and field studies to generate an understanding of Salmonella population genetics, microbial ecology, and transmission through commodities and their environments;
3) developing effective and affordable pre- and post-harvest interventions for use by commercial producers;
4) integrating data from these studies, data from monitoring programs, and machine learning to create decision support tools for each commodity, production stage, and company;
5) establishing communications and outreach efforts that allow stakeholders to provide feedback to further hone research priorities and facilitate implementation of the program.
I support the team by providing overall project management, data analytics, and data management, and serve as a project leader for Objective 4.
Check out the Salmonella Grand Challenge.
Doctoral Research: Understanding the amphibian-killing pathogen outside of the amphibian
Briggs Lab. UC Santa Barbara
Fungal pathogens are special-case pathogens which require unique methods for detection, modeling, and management. They have unique characteristics which set them apart from other pathogen groups (bacteria, viruses) and make it difficult to not only accurately detect the pathogen, but estimate its risk to focal hosts and predict disease dynamics. While different in many ways from other pathogen classes, fungal pathogens nonetheless can have similarly devastating impacts on forests, crops, and wildlife. Batrachochytrium dendrobatidis (Bd) is a primary example of a fungal pathogen which has caused devastating global declines and for which new methods are required for detection and prediction. In my PhD thesis, I explored three questions: (1) can we use alternative methods to detect Bd in the environment (i.e., not on amphibians); (2) can we predict Bd in amphibian populations using only information about the amphibian’s habitat; and (3) can Drosophila melanogaster be used as a model organism to study invertebrate-Bd dynamics? By answering these three questions, I aim to elucidate how Bd is able to cause variable responses in amphibian populations (i.e., epidemic vs. endemic states), and understand more about how Bd interacts with its environment as a generalist, fungal pathogen rather than an amphibian specialist.
Check out the Briggs Lab.
Check out the Briggs Lab.
Undergraduate Research: Mapping the spatial distribution of Batrachochytrium salamandrivorans
Zellmer Lab. Occidental College, Los Angeles, California
Which of these models of spread best predicts current Bsal presence and absence data?
Model selection with biased data (i.e., invasive species)
Bsal is thought to be endemic to East Asia, but it is currently causing declines in species naïve to it in Western Europe. Because of this, Bsal is not only an emergent pathogen but also an invasive species. A useful tool to predict the spread of invasive species, called Species Distribution Modeling (SDM), typically uses limited, presence/absence data and machine learning algorithms to identify suitable habitat for a species. Powerful, right? Only if the models can be tested against empirical data to ensure their accuracy! Many tools - model selection techniques - exist to test these models to ensure biological relevance. However, with an invasive species, data can be distributed based on spread dynamics (rather than where habitat is actually suitable), which can ultimately bias the model selection techniques as well. Therefore, is there a model selection technique that handles this biased data better than others? Can we use this technique to identify a model that can predict where Bsal will spread?
This question became the focus of my undergraduate research, and you can read more about it here!
Check out the Zellmer Lab.
Bsal is thought to be endemic to East Asia, but it is currently causing declines in species naïve to it in Western Europe. Because of this, Bsal is not only an emergent pathogen but also an invasive species. A useful tool to predict the spread of invasive species, called Species Distribution Modeling (SDM), typically uses limited, presence/absence data and machine learning algorithms to identify suitable habitat for a species. Powerful, right? Only if the models can be tested against empirical data to ensure their accuracy! Many tools - model selection techniques - exist to test these models to ensure biological relevance. However, with an invasive species, data can be distributed based on spread dynamics (rather than where habitat is actually suitable), which can ultimately bias the model selection techniques as well. Therefore, is there a model selection technique that handles this biased data better than others? Can we use this technique to identify a model that can predict where Bsal will spread?
This question became the focus of my undergraduate research, and you can read more about it here!
Check out the Zellmer Lab.
|
Chytrid survey with the USGS
While Bd is currently widespread across North America, Bsal has yet to make it over here. The USGS is currently swabbing salamanders around North America in the hopes of detecting Bsal as soon as it arrives. My lab and I aided in these efforts and helped to swab Slender Salamanders and California Newts across urban Los Angeles! |
A juvenile Batrachoseps nigriventris in urban Los Angeles
|
Publications
- Bosilevac, J.M., Katz, T.S., Manis, L.E., Rozier, L., Day, M. (2025). Using Pathogenic Escherichia coli Type III Secreted Effectors espK and espV as Markers to Reduce the Risk of Potentially Enterohemorrhagic Shiga Toxin-Producing Escherichia coli in Beef. Foods 14(3), 382. doi: 10.3390/foods14030382.
- King, D.A., Shackelford, S.D., Nonneman, D., Katz, T.S., Wheeler, T.L. (2025). Categorization of Beef Longissims Lumborum and Gluteus Medius Muscles Based on Metabolic Attributes is More Informative Than Muscle pH. Meat and Muscle Biology 9(1). doi: 10.22175/mmb.18388.
- Harhay, D.M., Brader, K.D., Katz, T.S., Harhay, G.P., Bono, J.L., Bosilevac, J.M., Wheeler, T.L. (2025). A novel approach for detecting Salmonella enterica strains frequently attributed to human illness – development and validation of the highly pathogenic Salmonella (HPS) multiplex PCR assay. Frontiers in Microbiology 15:1504621. doi: 10.3389/fmicb.2024.1504621.
- Byer, A.*, Nguyen, K.*, Katz, T.S., Chen, R., Briggs, C.J. (2024). Drosophila melanogaster is a possible vector of Batrachochytrium dendrobatidis. PLoS ONE 19(7):e0307833. doi: 10.1371/journal.pone.0307833.
- McMahon, T., Katz, T.S., Barnett, K.M., Hilgendorff, B. (2024). Centrifugation is an effective and inexpensive way to determine Batrachochytrium dendrobatidis quantity in clean water samples. Oecologia 205(3):437-443. doi: 10.1007/s00442-024-05604-0
- Le, M.,*, Meiman, A., Covey, A., Gole, A., Meng, M., Villa, N., Litvin, S., Katz, T.S., Deshmukh, R. (2024). Participation gap analysis among energy efficiency programs in California’s public sector. Energy Research and Social Science 114:103590. doi: 10.1016/j.erss.2024.103590.
- Bosilevac, J.M., Katz, T.S., Arthur, T.M., Kalchayanand, N., Wheeler, T.L. (2024). Proportions and Serogroups oof Enterohemorrhagic Shiga Toxin-producing Escherichia coli in Feces of Fed and Cull Beef and Cull Dairy Cattle At Harvest. Journal of Food Protection 100273. doi: 10.1016/j.jfp.2024.100273.
- Bosilevac J.M., Guragain M., Barkhouse D.A., Velez S.E., Katz T.S., Lu G. and Wang R. (2024) Impact of intense sanitization procedures on bacterial communities recovered from floor drains in pork processing plants. Frontiers in Microbiology 15:1379203. doi: 10.3389/fmicb.2024.1379203
- Katz, T.S., Harhay, D., Schmidt, J.W., Wheeler, T.L. (2024) Identifying a List of Salmonella Serotypes of Concern to Target for Reducing Risk of Salmonellosis. Frontiers in Microbiology 15:1307563. doi: 10.3389/fmicb.2024.1307563
- Zellmer A.J., Slezak P., Katz, T.S. (2020). Clearing up the Crystal Ball: Understanding Uncertainty in Future Climate Suitability Projections for Amphibians. Herpetologica 76(2): 108-120. doi: 10.1655/0018-0831-76.2.108.
- Katz, T.S., Zellmer, A.J. (2018). Comparison of model selection technique performance in predicting the spread of newly invasive species: a case study with Batrachochytrium salamandrivorans. Biological Invasions, 20: 2107-2119. doi: 10.1007/s10530-018-1690-7
