Genomic landscape in the human vaginal microbiome links to host geographics and genetics

The vaginal microbiota is essential for women’s health, yet its genomic diversity and host relevance remains incompletely characterized. Here we established a Chinese cohort with 10,281 vaginal metagenomes. We developed an extensive catalog of vaginal microbial genomes (VMG) by integrating 6,979 in-house metagenomes with 1,817 publicly available metagenomes and over 1,000 bacterial isolates. This catalog comprised 46,906 genomes from 913 prokaryotic species and 3,763 viral populations, with 23.8% of prokaryotes and 75.1% of viruses being not found in public reference databases. Leveraging VMG, we identified substantial intraspecies genomic and functional variations within the vaginal microbiome that display geographic specificity. Notably, this included the novel bacterium CAISGS01(ID u199), unique to Chinese samples with marked biosynthetic capabilities, as well as BVAB1, which exhibited distinct regional genomic variations in pathogenic potential. Moreover, by utilizing genome-resolved microbial profiles from the VMG, we conducted a metagenome-genome-wide association study involving 6,893 individuals, and identified 7 vaginal microbial taxa whose abundances were significantly associated with 31 host genomic loci, including a notable correlation between novel bacterium Prevotella (ID u35) and loci on 22p11.2 (beta = 1.51, p = 6.826 × 10 ^− 38 ). These associations were consistently and robustly replicated across three independent cohorts. In summary, our research provides a vital reference for future studies on the genotype-phenotype interplay within the human vaginal microbiome.