Genomic reconstruction of short-chain fatty acid production by the human gut microbiota

Short-chain fatty acids (SCFAs) including acetate, formate, propionate, and butyrate are the end products of dietary fiber and host glycan fermentation by the human gut microbiota (HGM). SCFAs produced in the column are of utmost importance for host physiology and health. Butyrate and propionate improve gut health and play a key role in the neuroendocrine and immune systems. Prediction of HGM metabolic potential is important for understanding the influence of diet and HGM-produced metabolites on human health. We conducted a detailed metabolic reconstruction of pathways for the synthesis of SCFAs and L- and D-lactate, as additional fermentation products, in a reference set of 2,856 bacterial genomes representing strains of >800 known HGM species. The reconstructed butyrate and propionate pathways included four and three pathway variants, respectively, that start from different metabolic precursors. Altogether, we identified 48 metabolic enzymes, including five alternative enzymes in propionate pathways, and propagated their occurrences across all studied genomes. We established genomic signatures for reconstructed pathways and classified genomes according to their simplified binary phenotypes encoding the ability (“1”) or inability (“0”) of a given organism to produce SCFAs. The resulting binary phenotypes combined into a binary phenotype matrix were used to assess the SCFA synthesis potential of HGM samples from several public metagenomic studies. We report baseline and variance for Community Phenotype Indices calculated for SCFAs production capabilities in 16S metagenomic samples of intestinal microbiota from two large national cohorts (American Gut Project, UK twins), the Hadza hunter-gatherers, and the young children cohort of infants with high-risk for type 1 diabetes. We further linked the predicted SCFA metabolic capabilities with available SCFA concentrations both for in vivo fecal samples and in vitro fermentation samples from previous studies. Finally, we analyzed differential representation of individual SCFA pathway genes across several WGS metagenomic datasets. The obtained collection of SCFA pathway genes and phenotypes enables the predictive metabolic phenotype profiling of HGM datasets and enhances the in silico methodology to study cross-feeding interactions in the gut microbiomes.