Objective To explore the pathogenesis of Crohn's disease by revealing the relationship between m6A methylation and Crohn's disease Methods The GEO (GENE EXPRESSION OMNIBUS) database was used to download the dataset GSE186582 on Crohn's disease, including standard tissue samples and Crohn's disease tissue samples, and the Expression of M6A-related genes in the calibrated dataset was obtained. Through the observation and comparison of the random forest tree method and machine learning method, it was determined that the random forest tree model could be used to screen the characteristic genes of diseases. Samples were divided into subtypes by the expression of m6A-related genes, and the relationship between different types and immune cells was analyzed and verified by principal component analysis. The expression of M6A-related genes and the relationship between the genotyped samples and immune cells were analyzed. We classified Crohn's disease according to the expression of differential genes, finally established the corresponding relationship between different types by Sankey diagram and analyzed the expression of Crohn's disease-related disease genes in two different types. Results By comparing the model construction methods, we found that the residual value of the random forest tree model method was low, and the area under the ROC curve was 1. Therefore, we chose the random forest tree method to construct the model and screen characteristic genes and found 11 methylation-related genes related to m6A in Crohn's disease, such as RBM15, YTHDF3 and RBM15B. According to the expression of 11 M6A-related genes, the samples were divided into two subtypes: activated B cells, immune B cells and MDSC (myeloid-derived inhibitory cells) expressed more than the B subtype (P value is less than 0). There was a significant positive correlation between the METTL3 gene, M6A recognition enzyme HNRNPA2B and activated CD4 + T cells. The expressions of activated B cells, MDSC and immune B cells in genotype B were significantly higher than those in genotype A (P < 0.05). Conclusion m6A modulators play an essential role in Crohn's disease, and the study of their patterns can guide future immunotherapy strategies for Crohn's disease

Objective To investigate the bioinformatics analysis methods of genes associated with colorectal cancer in ulcerative colitis. Methods We employed the intersection of the differential genes between UC and healthy controls, differential genes between UC dysplasia and UC, and the differential genes between UC dysplasia and healthy controls in GSE47908 to obtain overlapping genes and validated their accuracy in the TCGA dataset of COAD and GSE40967 to screen risk genes. The GSE110224/GSE113513 dataset of CODA, and the UC and COAD-related dataset GSE3629 were integrated for WGCNA analysis after normalizing the data. NOMO plot analysis was performed using the expression of overlapping genes of modular and risk genes in GSE47908 with UC dysplasia and UC. Results 1576 overlapping genes were detected after screening for differential genes, which were validated in the TCGA and GSE datasets of colorectal cancer to construct a prognostic model. It was found that all P-values were less than 0.05 after survival analysis and less than 0.05 for progression-free survival, and the area under the risk score curve of the ROC curve was 0.894, which could be more accurate as a predictor of patient prognostic indicators. Then, WGCNA analysis was performed on UC, COAD and healthy controls to obtain five modular genes and intersected with overlapping genes to obtain 490 overlapping genes, and NOMO plotting by the LASSO algorithm to obtain seven key genes to predict the risk score of UC progression to COAD. Conclusion We screened seven gene indicators that could be used as key biomarkers of colorectal cancer susceptibility in patients with ulcerative colitis.