Recipient Organization
UNIV OF MARYLAND
(N/A)
COLLEGE PARK,MD 20742
Performing Department
(N/A)
Non Technical Summary
Foodborne illness impacts nearly 50 million people in the United States per year, causing severe economic damages and healthcare burdens. There is an urgent need for more effective strategies to manage pathogen transmission across food and agricultural systems. This proposal aims to definethe spatial and temporal distributions of expansive genomic diversity of foodborne pathogens by integrating open-source molecular data with artificial intelligence (AI) and machine learning (ML) technologies, in order to inform practical applications that will promote food safety. Focusing on a variety of pathogens that present unique risks (i.e.,Campylobacter jejuni,Cronobacter sakazakii, andListeria monocytogenes), we will build and validate a computational workflow to understand how selective pressures across food chains drive bacterial environmentaltropisms (e.g., genes underlying persistence, antimicrobial resistance, or virulence linked to specificsource types). Leveraging publicly available genomic and phenotypic data will enable high-throughput identification of new molecular markers associated with foodborne pathogen transmission and evolution. These data will further serve as key reference material to improve strategies for pathogen molecular surveillance with metagenomics. Overall, the analytical framework that wedevelop and promote for use by the food safety research community will have applications that extend to broader foodborne pathogen groups and emerging food safety concerns.
Animal Health Component
30%
Research Effort Categories
Basic
70%
Applied
30%
Developmental
(N/A)
Goals / Objectives
Our proposal aims to integrate AI and ML with open-source molecular data to define the spatial and temporal distributions of expansive genomic diversity of foodborne pathogens. The data-analytic workflows that we buildwill have innovative technological applications for food and agricultural industries to support food safety, nutrition, and health. Objective 1:Establish systematic pipeline to elucidate the biogeography of genomic diversity of foodborne pathogens.Objective 2:Benchmark metagenomic tools for foodborne pathogen detection in simulated microbial communities.These independent, yet interrelated objectives will test ouroverarching hypothesisthat environmental stressors along the agricultural continuum to consumers impose selective pressures that drive adaptive responses of diverse foodborne pathogens -C. jejuni,C. sakazakii, andL. monocytogenes- and the emergence of key genes and pathways (e.g., persistence, antimicrobial resistance, virulence) that hold potential for informing novel applications to promote food safety.
Project Methods
Objective 1will establish a reproducible bioinformatics pipeline to annotate and compare pangenomes of critical foodborne pathogens that represent different bacterial lifestyles, survival strategies, and foodborne illness implications (C. jejuni, C. sakazakii,andL. monocytogenes). Leveraging publicly available resources, including >100,000 open-source genome assemblies archived in 'NCBI Pathogen Detection,' will enable high-throughput identification of new molecular markers associated with pathogen transmission and clinical health risks. Genome assembly accessions will be sourced from bacterial isolate metadata, focusing on isolates with assigned source categories (food, environment, and clinical), origin location, and year. Quality control programs will filter and select forhigh quality assemblies, which will be used in downstream pangenome analysis to characterize the comprehensive genetic diversity of respective target species' groups. Additional tools will be employed for feature identification within pangenomes (e.g., antimicrobial resistance genes, virulence genes, metabolic pathways). Machine learning models will be used to uncover associations between the metadata, notably source type, and genomic features.Key genes and pathways putatively associated with environmental tropisms with food safety implications will be identified across geographic and temporal scales for an extensive set of food commodities.These data will further serve as reference material to optimize metagenomic frameworks for pathogen molecular surveillance inObjective 2. Expanding on acquired knowledge of the biogeography of genomic diversity ofC. jejuni,C. sakazakii, andL. monocytogenes, we will explore how strain-level variation impacts detection potential for microbiome profiling with metagenomics. Traditional metagenomic tools will be benchmarked for efficacy in predicting abundances of the respective pathogens in simulated microbial communities, testing effects of pathogen concentration and strain-level variation (i.e., different lineages or clades of each species) on metagenomic predictions. Overall, the AI-based pangenome and metagenome workflows that we develop will have broad applications in future research that extends to additional foodborne pathogen groups and emerging food safety concerns.