USDA Grant Backs UMD Research Fighting Foodborne Illnesses

One thing is easy to predict about food contamination: With an estimated one in six Americans sickened each year, according to the Centers for Disease Control and Prevention, sooner or later everyone will face it.

But it’s harder to understand how the germs that cause some of the most troubling foodborne illnesses—E. coli, salmonella and listeria—survive, proliferate and occasionally cause outbreaks throughout the food system.

Research by Professor Abani Pradhan in the Department of Nutrition and Food Science and the Center for Food Safety and Security Systems at the University of Maryland will help fill in the missing information. The U.S. Department of Agriculture’s National Institute of Food and Agriculture recently awarded Pradhan $591,000 to develop new tools using genomics and machine learning to better predict the conditions that can lead to foodborne illness outbreaks.

“The whole genome sequencing and metagenomics data are mostly available or can be generated on all these pathogens that cause these very serious foodborne illnesses,” said Pradhan. “But there is really not a simple tool for using it to improve food safety.”

Pradhan and his team will use machine learning to analyze the genomes of salmonella, listeria and E.coli from publicly available databases to find genetic indicators—like specific genes, mutations, or higher or lower levels of gene expression—that help them persist in the environment, perhaps by resisting cleaning agents, evading human immune responses or surviving various temperature and moisture conditions.

Then the team will evaluate the environmental conditions surrounding outbreaks of these diseases from facilities where poultry and leafy green vegetables are raised. Researchers will look at factors like temperature, moisture and the microbiome in food processing environments, such as the soil where leafy greens grow, or the water used for irrigation.

By combining all of this information with the genomic data and analyzing it with machine learning, the team expects to find patterns that can more accurately predict what conditions and circumstances lead to outbreaks.

Their final step will be to test their methods by comparing the predictions of their machine learning model in real-world settings, whether a laboratory greenhouse or farm environment. Pradhan will partner with collaborators from the USDA and the U.S. Food and Drug Administration to evaluate the work.

Pradhan’s ultimate goal is to develop a digital dashboard that can help decision-makers predict and track the potential persistence or emergence of pathogens and take appropriate actions to prevent outbreaks.