identification of immune-related diagnostic genes and subtypes of active ulcerative colitis

Ulcerative colitis (uc) is a chronic immune-mediated disease with a growing global burden. in this study, bioinformatics analysis and machine learning methods were employed to screen the potential immune microenvironment-related feature genes. we identified 4 immune-related genes that are consistently dysregulated in uc and correlate with immune infiltration, including apobec3b (apolipoprotein b mrna editing enzyme catalytic subunit 3b), cxcl11, pla2g2a, and tmem173. their diagnostic performance was verified in an external cohort and in our clinical samples. then, the proportion of ambiguous clustering (pac) successfully classified uc patients into 2 molecular subtypes, including subtype 1 (metabolism-related subtype) and subtype 2 (immune-related subtype). the single sample gene set enrichment analysis (ssgsea) algorithm revealed that subtype 2, with a higher score, of the majority of immune cells presented a worse inflammatory response. in addition, we assessed scores of partial novel drugs querying the cmap database and found that the efficacy of clinical small-molecule compounds presented different results across uc subtypes. these findings identify biomarkers, establish a concise immune-based classification of uc, and support subtype-guided therapy.