Identification of Inflammation-Related Diagnostic Biomarker and Molecular Subtypes in Ulcerative Colitis Based on Machine Learning.

backgroundUlcerative colitis (UC) is a chronic and recurrent form of inflammatory bowel disease, primarily involving sustained inflammation of the colonic mucosa. The inflammatory process plays a pivotal role in the pathogenesis and progression of the disorder. Therefore, we conducted this research to explore potential biomarkers and classify molecular subtypes with clinical relevance, which enhances diagnostic accuracy and informs personalized therapeutic strategies.

methodsTranscriptomic information was retrieved via the Gene Expression Omnibus (GEO) database. After merging the raw data, differential gene expression analysis was carried out, accompanied by a suite of bioinformatics tools. Gene Ontology (GO) annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed. To screen feature genes, we employed both least absolute shrinkage and selection operator (LASSO) regression and random forest (RF), representing distinct approaches within machine learning. The effectiveness of the selected genes in distinguishing disease status was measured using receiver operating characteristic (ROC) curve analysis, and a nomogram was constructed. We employed single-sample gene set enrichment analysis (ssGSEA) to quantify the enrichment of immune-related signatures at the sample level, while gene set enrichment analysis (GSEA) was conducted to explore key signaling pathways. In addition, drug sensitivity analysis was performed to predict potential therapeutic responses. Ultimately, qRT-PCR analysis was performed to validate the expression patterns of the selected marker genes.

resultsTwo feature genes (CCL11 and MMP1) were identified as diagnostic biomarkers, both demonstrating strong diagnostic performance in ROC curve. A comprehensive nomogram was constructed and shown to possess considerable clinical utility. In addition, significant alterations in immune cell infiltration were identified, and key biological pathways were revealed through GSEA.

conclusionThis study identified CCL11 and MMP1 as inflammation-related diagnostic biomarkers for UC through integrated machine learning approaches, offering potential directions aiming at personalized diagnosis and therapeutic intervention.