Dissecting Gene-Microbe-Exercise Interactions in Determining Host Phenotypes

Abstract

Host genetics, gut microbiome, and physical activity are all important factors in determining human health. Clinical chemistries are commonly used in research and medicine to understand, diagnose, or treat disease. Previous metabolomic research has shown that there are gene, microbe, and hybrid associations with blood metabolites. Here, I investigate the possibility that physical activity moderates the relationship between genome-driven or microbiome-driven metabolites genetics and clinical chemistries. Using the Arivale cohort, a large-scale wellness study that includes health and molecular phenotypic data, I examine monthly average distance traveled (determined by FitBit tracking) as an effect modifier in this relationship. Through ordinary least squares linear regression, I found 174 metabolites that are associated with activity, of which 35 were genetic-only, 40 were microbe-only, and 99 were under hybrid control. Of these, 7 metabolites were chosen to undergo moderation analysis which uncovered 5 metabolite-chemistry pairs where activity was a significant effect modifier (p-value < 0.05 after FDR correction). Each unique pair tells a story, which could help explain the underlying mechanisms of physical activity on health.