Host-microbe interactions are crucial for normal physiological and immune system development and are implicated in a wide variety of diseases, including inflammatory bowel disease (IBD), obesity, colorectal cancer (CRC), and type 2 diabetes (T2D). Despite large-scale case-control studies aimed at identifying microbial taxa or specific genes involved in pathogeneses, the mechanisms linking them to disease have thus far remained elusive. To better identify potential mechanisms linking human-associated bacteria with host health, we leveraged publicly-available interspecies protein-protein interaction (PPI) data to identify clusters of homologous microbiome-derived proteins that bind human proteins. By detecting humaninteracting bacterial genes in metagenomic case-control microbiome studies and applying a tailored machine learning algorithm, we are able to identify bacterial-human PPIs strongly linked with disease. In 9 independent case studies, we discover the microbiome broadly targets human immune, oncogenic, apoptotic, and endocrine signaling pathways, among others. This host-centric analysis strategy illuminates human pathways targeted by the commensal microbiota, provides a mechanistic hypothesis-generating platform for any metagenomics cohort study, and extensively annotates bacterial proteins with novel hostrelevant functions.