Location: Genomics and Bioinformatics Research
Project Number: 6066-21310-006-041-S
Project Type: Non-Assistance Cooperative Agreement
Start Date: Sep 15, 2025
End Date: Dec 31, 2028
Objective:
Salmonella bacteria sicken 1.4 million Americans annually and cost the US economy $3.7 billion. Some types of Salmonella make people really sick, while others are less harmful. A key challenge to prevention is that we cannot yet predict how pathogenic a new Salmonella strain is. This work will develop the data and machine learning models (a type of artificial intelligence) that can predict the pathogenicity of a new Salmonella strain. This tool will help producers and food processors target the most dangerous Salmonella strains for control.
Approach:
Figuring out why some Salmonella strains are more dangerous than others is tricky because each strain can have hundreds of virulence factors that contribute to its ability to cause disease. This project builds computer models that can not only predict which strains are dangerous but also explain why. This project combines lab experiments with machine learning to do this. As part of this project, the cooperator has created a collection with diverse Salmonella types. The next step is to do the same with older strains of Salmonella that existed before modern food safety measures were in place, giving a broader view over time. To test how harmful these bacteria are, the cooperator will see how well they stick to and invade immune cells—specifically macrophages—and how well they reproduce inside them. These collections will be used by the cooperator to quickly test other traits that affect how Salmonella behaves in food, animals, and people. ARS has characterized the genetic variation from over 700,000 Salmonella strains. By combining this information with the lab data from carefully selected strains, the project will build models that make better predictions about how dangerous a new strain might be just from looking at its DNA. Ultimately, this work could help identify and target risky Salmonella strains for management before they create outbreaks, keeping people safer and food systems more secure.