(590d) Metagenome Based Metabolic Network Analysis of Human Microbiome

Chen, Y. - Presenter, University of Michigan
Lin, X. N. - Presenter, University of Michigan, Ann Arbor

The human body harbors various microbial communities (i.e. human microbiome) that are believed to affect the host health but remain largely enigmatic. Recent years have seen rapidly increasing use of metagenomic sequencing for study of natural microbial communities and the human microbiome has become one of the latest targets of such metagenomic investigation. In this work, we set out to make use of these large-scale metagenomic datasets to elucidate the metabolic functions of the human microbiome and the microbe-microbe-host interactions, by developing and utilizing a metabolic network modeling framework.

Using a gut microbiota metagenomic dataset generated from sampling of 124 European individuals, we reconstructed in silico metabolic networks of various phylogenetic groups at different taxonomic levels. To represent variations among samples (people), the frequency (discovery rate) of each metabolic reaction was calculated prior to network reconstruction. Then minimal metabolic networks for biomass synthesis were generated for different phylogenetic groups, each consisting of 1,300-2,200 reactions. Furthermore, minimal reaction sets for the synthesis of individual biomass components were examined based on the entire dataset and their discovery rates in different phylogenetic groups were compared. Our results suggested distribution of certain metabolic functions, such as biosynthesis of short-chain fatty acids and vitamins, in the microbial community.

To directly analyze the interactions among microbiome members, we also developed multiple-compartment network models that consider the most abundant groups at the same time.  It was found that many metabolic interactions potentially exist among the microbial groups, e.g. between Bacteroidetes, Firmicutes and Proteobacteria. These results indicated that metabolic cooperation among microbes is common and might be important for sustaining robust human microbiota.