Genomic Model Identifies Chicken and Vegetables as Key Sources of Salmonella Risk

TL/DR –

A recent study in the U.S. used genomic sequencing and machine learning to trace the primary food sources of Salmonella infections. Using a dataset of 18,661 Salmonella isolates from food and animal samples, researchers found chicken (31%), vegetables (13%), turkey (12%), and pork (11%) to be the main sources. When applied to human infections, the model attributed 34% of illnesses to chicken and 30% to vegetables, accounting for almost two-thirds of infections.

A groundbreaking genomic model uncovers the real risk of Salmonella— pinpointing chicken and vegetables as prime sources and revolutionizing the fight against foodborne illness.

Study: Attribution of Salmonella enterica to Food Sources by Using Whole-Genome Sequencing Data. Credit: nobeastsofierce / Shutterstock

A recent study in the journal Emerging Infectious Diseases utilized genome sequencing and machine learning to pinpoint the primary food sources of human Salmonella infections in the US.

Study Insights

Salmonella enterica infections annually result in about 1.35 million illnesses in the US. Common sources are contaminated food, water, animals and soil. With Whole-Genome Sequencing (WGS), we can better understand Salmonella transmission pathways. This knowledge is vital for improving food safety regulations and prevention tactics.

The researchers analyzed 18,661 Salmonella isolates from food and animal samples. These were categorized into 15 distinct food groups. To balance the dataset, 50% of chicken isolates were randomly selected and inverse class weighting was used to correct imbalances.

To identify infections in humans, 6,470 Salmonella isolates were collected from individuals who had not traveled internationally. The team used SPAdes software to compile genetic data and applied whole-genome multilocus sequence typing (wgMLST) to both food-derived and human isolates. They trained a Random Forest machine learning algorithm on isolates with known sources. The optimized model predicted infection sources for human cases with over 50% probability.

Study Outcome

The model, trained on genomic data, identified chicken (31%), vegetables (13%), turkey (12%), and pork (11%) as the primary Salmonella sources. When applied to human infections, the model attributed 34% of illnesses to chicken and 30% to vegetables, collectively accounting for roughly 73% of confirmed sources.

The accuracy of the model was high, especially for chicken (97%), vegetables (82%), turkey (88%), pork (83%), and beef (77%). The findings emphasize the requirement for interventions targeting poultry and fresh produce to reduce the Salmonella burden. For better accuracy, it’s suggested to expand the dataset with more diverse non-chicken isolates and other non-food sources. Broader, nationwide data collection is also recommended.

Summary

The study showcases the effectiveness of WGS coupled with a Random Forest machine learning algorithm to accurately identify Salmonella food sources in the US. Chicken and vegetables are the prime contributors, reinforcing the need for targeted regulatory and health strategies. This genomic approach offers valuable insights for food safety policy, surveillance, and outbreak management. Further research should broaden sample diversity, expand geographic representation, and include non-food sources to enhance the model’s precision.

—
Read More Health & Wellness News ; US News