Gene-centric metagenomic analyses reveal microbiome functional insights into diseases

The microbiome encodes millions of genes; however, understanding their role in human health remains challenging. Here, we developed MetaGEAR, a gene-centric analysis framework for metagenomic data. MetaGEAR constructs cohort-specific databases for efficient retrieval of gene annotations, abundances, and co-localization, while providing enhanced taxonomic resolution by integrating reference- and assembly-based approaches. This is combined with a Metagenomic Assembled Graph (MAGraph) capturing gene neighborhood information. Using MetaGEAR, we built a multi-cohort database comprising >33 million gene families to investigate microbiome functionality across 24 cohorts of inflammatory bowel disease, colorectal cancer, and healthy populations and identified disease signature genes. Furthermore, the MAGraph revealed mobile genetic elements acting as hubs for tetracycline resistance spreading among Enterococcus , Streptococcus , and Veillonella species. Also, a duplicated nitrate reduction operon in Klebsiella pneumoniae was linked to differential gene expression under stress and virulence-inducing conditions. In summary, gene-centric metagenomic analyses reveal important insights into microbiome functionality in diseases.