〔Abstract〕 To explore the causal relationship between 46 phenotypes ( including 15 seropositive case- control phenotypes and 31 quantitative antibody-measurement phenotypes) and ulcerative colitis( UC) using two- sample bidirectional Mendelian randomization( TSMR) . Methods Single nucleotide polymorphisms ( SNPs) sig- nificantly associated with the relative abundance of the 46 antibody sera were extracted as instrumental variables ac- cording to preset thresholds . Summary statistics for UC were obtained from the OPEN GWAS database ( n = 47 745) . MR-Egger regression , weighted median method ( WME) , inverse variance weighting ( IVW) , the simple mode method (SM) , and weighted multitude method (WM) were used to estimate the causal relationship between antibody levels and UC , primarily using the IVW method . The results were assessed according to the effect indica- tor dominance ratios (OR) and 95% confidence intervals (CI) . Sensitivity analysis , heterogeneity test , gene plei- otropy test , and outlier test (MR-PRESSO) were combined to verify the stability and reliability of the results , and the causal association study was performed again using reverse Mendelian randomization(MR) . Results IVW re- sults showed that Epstein-Barr( EB) virus EA-D antibody levels ( OR = 0. 806 , 95% CI = 0. 693 - 0. 939 , P < 0. 01) , Epstein-Barr virus EBNA-1 antibody levels ( OR = 1 . 870% , 95% CI = 1 . 480 - 2. 360 , P < 0. 000 1) , Anti-polyomavirus 2 IgG seropositivity (OR = 0. 570 , 95% CI = 0. 435 - 0. 746 , P < 0. 000 1) were associated with UC . The inverse MR analysis revealed a causal effect on anti-polyomavirus 2 IgG seropositivity , and none of the a- bove revealed genetic pleiotropy or significant heterogeneity of IVs . Conclusion EB virus EBNA-1 antibody levels are positively associated with the risk of UC , while EB virus EA-D antibody levels and anti-polyomavirus 2 IgG se- ropositivity are negatively associated with the risk of UC , indicating that they are protective factors for UC .