An expansive animal gut microbiome dataset elucidates major compositional shifts across bilaterian evolution

Animal gut microbiomes provide key physiological functions and are critical for host health. They vary dramatically across the animal kingdom, and are shaped by factors including host diet, evolutionary history and environment. However, analyses of gut microbiomes spanning the entire metazoan clade are lacking, limiting our understanding of the fundamental principles governing gut microbiomes. Here we present the Gut Microbiome Tree of Life (GMToL), a curated 16S amplicon dataset of 17,366 samples from 1,553 host species across 26 host classes from 284 studies, enabling analysis of large-scale evolutionary trends. Using ancestral state reconstruction, we provide a critical baseline calculation of major compositional shifts in gut microbiomes throughout animal evolution. We show that the ancestral animal gut was likely dominated by Pseudomonadota. A major shift to Bacteroidota occurred during the evolution of tetrapods, followed by the emergence of Bacillota-dominated guts in mammals and birds. We identify conserved core gut microbes and demonstrate how GMToL can be leveraged to contextualize the evolutionary history of specific microbial taxa. Ultimately, this framework enables the predictive mapping of microbial symbionts across uncharacterized host lineages, and establishes a quantitative baseline for comparative microbiome research at scale.