Michael Henson talk March 2019

Metabolic modeling of the complex microbial communities responsible for promoting health and causing disease in human hosts has emerged as an important in silico tool for investigating community stability, robustness and function. While yielding important insights into community behavior, most microbiota models developed to date have been limited with respect to the number of species included and/or the metabolic interactions allowed. In this talk, our recent work on developing large-scale models for bacterial communities resident in the human gut and cystic fibrosis lung is discussed. The human gut microbiota model includes 20 species representing common genera determined through 16S mRNA sequencing of stool samples. The community model is used to investigate the tradeoff between community growth and diversity. Compared to maximal growth, suboptimal growth solutions are characterized by higher species diversity and more balanced synthesis of health-promoting short-chain fatty acids. These predictions are consistent with known characteristics of healthy gut communities, suggesting that simulated suboptimal growth represents a “healthy” state and simulated maximal growth represents a “dysbiosis” state such as inflammatory bowel disease. A 17 species model of the cystic fibrosis (CF) microbiota is developed from 16S sequence data reporting the most abundant genera in 75 sputum samples collected from 46 adult CF patients. By performing Monte Carlo simulations with randomized community uptake rates to simulate sample-to-sample heterogeneity, we are able to rationalize the frequent domination of the community by Pseudomonas and Streptococcus as well as infrequent domination by Burkholderia, Enterobacteriaceae or Achromobacter. The co-domination of Pseudomonas and Streptococcus is predicted to be driven by crossfeeding of specific metabolic byproducts and amino acids.