8515 – Probing Host-Gut Microbiome Protein-Protein Interactions to Identify Therapeutic Targets and Diagnostic Biomarkers of Disease

Using a novel machine-based learning approach, prokaryote-humancell interactions in the human microbiome has revealed numerous noveltherapeutic targets and biomarkers of disease. The pathogenesisof many human diseases has been linked to interactions between the microbiome andits human host in the form of protein-protein interactions (PPIs). Theidentification of such interactions can improve current diagnostic andtherapeutic tools for microbiome-associated diseases, including type 2diabetes, colorectal cancer, inflammatory bowel disease, and obesity. A majorsource of PPIs is located in the gut, where over 3 million microbiome genes areexpressed in proximity of human cells. Yet, current in silico PPImodeling approaches have not been applied to the human gut microbiome. By leveragingexperimentally verified interspecies PPIs, Cornell researchers have devised ahomology-based platform that models millions of human-microbiome PPIs. Theresearchers use this database in conjunction with a machine learning algorithm toidentify disease-specific PPIs based on microbiome metagenomic reads fromseveral case control studies. Phillip Owh po62@cornell.edu 1-607-254-4508