Gut microbiome-disease associations revealed by large-scale Mendelian randomization analyses

Abstract Background The gut microbiome has been associated with diverse human diseases, but systematic genetic evidence for potential causal relevance, disease-system clustering, and circulating metabolic pathways remains limited. Methods We integrated four gut microbiome genome-wide association study (GWAS) datasets with 4,934 disease GWAS outcomes. Two-sample Mendelian randomization (MR), using the inverse-variance weighted (IVW) method as the primary analysis, was performed to evaluate potential causal associations between gut microbial traits and diseases. Multiple sensitivity analyses and reverse MR were conducted to assess robustness and potential reverse causality. Hypergeometric tests were used to examine disease-system enrichment of microbial lineages with significant MR associations. For significant MR associations, Bayesian colocalization analysis was performed to assess shared local genetic signals. Circulating metabolite GWAS data were further incorporated into two-step MR mediation analysis to explore candidate metabolic pathways. Results We identified 1,943 significant associations between 630 gut microbial traits and 925 disease outcomes across multiple disease systems. Enrichment analysis revealed system-specific clustering patterns, including enrichment of Burkholderiales-related taxa in circulatory diseases, Bacteroidota-related taxa in mental, behavioural, or neurodevelopmental disorders, Eisenbergiella -related taxa in musculoskeletal diseases, and Faecalicatena -related taxa in endocrine and metabolic diseases. Colocalization analysis identified 58 colocalization signals with supportive evidence of shared local genetic signals, involving 34 microbial traits, 45 disease outcomes, and 33distinct genomic windows. Mediation analysis highlighted triglycerides in intermediate-density lipoprotein (IDL-TG) as a frequent candidate mediator, particularly linking Burkholderiales-related taxa to ischaemic cardiovascular diseases. Conclusions This large-scale genetic study provides a systematic map of potential causal associations between the gut microbiome and multi-system diseases. By integrating MR, disease-system enrichment, colocalization, and mediation analyses, our findings suggest that specific gut microbial lineages may contribute to disease susceptibility through shared genetic architecture and circulating metabolic pathways. In particular, the Burkholderiales-IDL-TG-ischaemic cardiovascular disease axis may represent a candidate microbiome-metabolism-disease pathway that warrants further validation in experimental and clinical studies.