Abstract
Alterations in the gut microbiota have been associated with a variety of medical conditions such as obesity, Crohn’s disease and diabetes. However, establishing the causality between the microbial composition and disease remains a challenge. We introduce a strategy based on metabolic models of complete microbial gut communities in order to derive the particular metabolic consequences of the microbial composition for the diabetic gut in a balanced cohort of 186 individuals. By using a heuristic optimization approach based on L2 regularization we were able to obtain a unique set of realistic growth rates that allows growth for the majority of observed taxa in a sample. We also integrated various additional constraints such as diet and the measured abundances of microbial species to derive the resulting metabolic alterations for individual metagenomic samples. In particular, we show that growth rates vary greatly across samples and that there exists a network of bacteria implicated in health and disease that mutually influence each others growth rates. Studying individual exchange fluxes between the microbiota and the gut lumen we observed that consumption of metabolites by the microbiota follows a niche structure whereas production of short chain fatty acids by the microbiota was highly sample-specific and was altered in type 2 diabetes and restored after metformin treatment in samples from danish individuals. Additionally, we found that production of butyrate could not be easily influenced by single-target interventions.