Gut Microbiota, Inflammatory Bowel Disease and the Prioritization of Host Effector Genes: A Two-Sample Mendelian...

Gut microbial dysbiosis is common in inflammatory bowel disease (IBD), but whether specific taxa are causal, and through which host genes, remains unclear. We integrated two-sample Mendelian randomization (MR), replication, cis-eQTL colocalization, and single-cell mapping to identify microbial taxa with genetically predicted associations and candidate host mediators. Instruments for 211 taxa from MiBioGen (p < 1×10^-5, r² < 0.001) were used in two-sample MR (inverse-variance weighted IVW primary; sensitivity: Egger, weighted median, MR-PRESSO) against IBD, Crohn's disease (CD), and ulcerative colitis (UC) in the International IBD Genetics Consortium (IIBDGC), with false discovery rate (FDR) correction. Replication was conducted in FinnGen R12; gene prioritization was performed via cis-SMR/coloc.abf using GTEx v8/CAGE eQTLs, and localization in 11,175 colonic cells. MR identified 15 significant associations (FDR < 0.05) across eight features; applying a pre-specified heterogeneity-based rule, 13 across seven independent features were retained as primary associations (strongest NB1n → CD, OR 1.32) and two were downgraded to suggestive; at the independent-signal level, three of 11 associations were replicated and seven were direction-concordant in FinnGen. Colocalization supported shared variants at SENP7 (blood, CD PP.H4 = 0.86, IBD = 0.70) and ZBTB11-AS1 (colon, IBD = 0.76, CD = 0.74); single-cell analysis mapped SENP7 to T/NK, and ZBTB11-AS1/VIPR1 to epithelial/goblet. Replicated MR and colocalization prioritized SENP7 (whole-blood eQTL; minor T/NK-cell detection) and ZBTB11-AS1 (colon epithelium) as candidate effectors, though the latter lacks functional characterization.