Development of a Genome Scale Model-Based Computational Framework to Predict Metabolic and Nutritional Drivers of Gut Microbiome Community Structure

Abstract

Emergent human gut microbiome structures impact health outcomes. However, what drives the assembly and interactions between microbes that enable such structures remains underexplored. Metabolic characterisation of individual gut strains in differing nutrient environments can contribute to improved understanding of microbial community assemblage, and potentially lead to rational design of health promoting interventions. Here, a computational high throughput genome scale modelling platform, termed GEMNAST, is developed to allow inference of the metabolic attributes and nutritional requirements of gut strains. GEMNAST is applied to characterise 816 gut strains, for which curated genome-scale models are available, in terms of three prominent nutritional dimensions: human vitamins, amino acids and carbohydrates. Doing so reveals contrasting metabolic patterns between prominent gut microbiome taxa. GEMNAST-generated data is further applied to infer nutritional interactions within emergent gut microbiome structures in a human cohort. Results show that different nutrients are preferentially exchanged in communities depending on their dominant taxa. Notably, modular-level organisation occurs around strains capable of degrading inulin in Ruminococcus dominated communities, or mucin and complex fibres in Bacteroides dominated communities. Finally, a predictive prototype that utilises GEMNAST-generated data, termed GEMNAST-predict, was developed as a proof-of principle for the design of rational health interventions. Analysis of human microbiome data suggests that prebiotic fibre intervention responses are likely driven by primary fibre degraders, based on their abundance. This may be linked to the greater levels of functional redundancy in the microbiome. Together, this body of work provides mechanistic grounding to gut microbiome stability and to co-abundance based observations, a fundamental step towards understanding emergent processes that influence health outcomes.