Integrative Analysis of the Mouse Cecal Microbiome Across Diet, Age, and Metabolic State in the Diverse BXD Population

The gut microbiota both adapts to, and shapes, the metabolic state of individuals. This bidirectional relationship is mediated via circulating metabolites and gene regulatory networks and interacts with many organs, including by the gut-brain axis. Here, we have processed the cecum from 232 mice from our recent aging colony across age (6-24 months), diet (chow or high fat), and genetics (43 BXD strains) and sequenced their metagenome, metatranscriptome, and cecal transcriptome. We quantify changes in over 300 species caused by interactions between diet, age, and genetic background. Traditional bioinformatics approaches linked particular microbes to observed phenotypes, while newer machine learning models based microbial clusters accurately predicted host outcomes, including individual body weight (AUC = 0.92) and chronological age (AUC = 0.84). This was further enhanced by a compact 10-feature multi-omics model, combining our microbiome data with prior liver expression data to increase chronological age AUC to 0.95. Mechanistically, integrative network analyses identified dozens of significant links between particular bacterial taxa and gene expression, such as a strong negative correlation between host Ido1 expression and short-chain fatty acid (SCFA)-producing Lachnospiraceae , indicating dietary fat can modulate host tryptophan metabolism via microbiota shifts. Moreover, as our study uses inbred mice sampled across time, we have identified signature sets of taxonomies that provide excellent predictive value for future metabolic outcomes driven by metabolic networks connecting the microbiome to the host organism’s tissues (here, cecum and liver from the same mice). By better understanding the gut-liver axis, we can understand the cellular etiologies of metabolic disease and identify earlier, personalized diagnostic biomarkers attuned to the genetic background and environmental state of the individual.