Prokaryotic genomes for 2058 new species assembled from globally distributed human gut metagenomes. Global Human Gut MAG (HGM) Dataset

Largely due to challenges cultivating microbes under laboratory conditions, many species in the human gut microbiome do not have a genome sequence. To address this problem, we reconstructed 60,664 prokaryotic draft genomes from 3,810 fecal metagenomes from geographically and phenotypically diverse human subjects. These genomes provide reference points for 2,058 previously unknown uncultured species and nearly double the phylogenetic diversity of sequenced gut Bacteria. New species comprise 33% of species richness and 28% of species abundance per individual on average and are more common in humans living rural lifestyles. A metagenome-wide association study of ten human diseases revealed thousands of new species-level biomarkers, particularly for colorectal cancer, ankylosing spondylitis, and obesity. Our analysis also uncovered novel genomic and ecological features of uncultured species, including reduced genome size, metabolic auxotrophies, and increased richness in high-diversity communities. We anticipate that this dataset and associated tools will drive new culturing efforts and enable further discoveries from the human gut microbiome.