Michael Skinnider, M.D., Ph.D.

Research

My research group will train artificial intelligence to characterize the “dark matter” of the human metabolome. The human body produces thousands of metabolites: small molecules that are generated and consumed by metabolic reactions and can serve as molecular biomarkers for disease diagnosis and treatment. However, the identity of most of these molecules—which together make up the “metabolome”—remains unknown. Previously, my research leveraged the power of AI to discover dozens of novel metabolites. Now, my lab will direct our AI model to optimize its own training, allowing us to prioritize collecting data that will fill the gaps in our model’s understanding of the human metabolome and then decode the structures and identities of these missing molecules. We will then point this AI model toward metabolites that are regulated by human genetic variation —work that will elucidate novel metabolic pathways and highlight previously unknown molecular mechanisms of disease.