Analysis of diagnostic genes and molecular mechanisms of Crohn's disease and colon cancer based on machine learning algorithms.

Crohn's disease (CD) is a chronic inflammatory bowel condition, and colon adenocarcinoma (COAD), as one of the most prevalent malignant tumors of the digestive tract, has been indicated by research to have a close association with CD. This study employs bioinformatics techniques to uncover the potential molecular links between CD and COAD. In this study, two data series related to CD were identified from the Gene Expression Omnibus (GEO) database under specific criteria, and relevant COAD gene data were obtained from The Cancer Genome Atlas (TCGA). Weighted Gene Co-expression Network Analysis (WGCNA), differentially expressed genes (DEGs), and protein-protein interaction (PPI) network analysis were conducted. A diagnostic model was established using machine learning. The accuracy of the diagnosis was validated using methods such as the construction of Receiver Operating Characteristic (ROC) curves and nomograms. Gene Set Enrichment Analysis (GSEA) was also employed to enrich the relevant pathways and biological processes. This study identified three genes through machine learning selection: DPEP1, MMP3, and MMP13. The ROC curves demonstrated that the machine learning model constructed with these three genes has a high level of accuracy, confirming their potential as biomarkers. Furthermore, GSEA elucidated that the pathways associated with these three key genes are closely related to cytokines and other factors. This study has identified key biomarker genes for CD and COAD: DPEP1, MMP3, and MMP13, providing additional molecular mechanism associations between the two diseases. It also offers more connections and pathways for reference regarding the progression of CD to COAD.