Comprehensive identification of three key genes and three key DNA methylation CpG sites associated with prognosis and prediction in colon adenocarcinoma.

Colon adenocarcinoma (COAD) is a common cancer with high incidence and mortality rates worldwide. DNA methylation has a significant impact on the occurrence and development of COAD. However, the regulatory and prognostic effects of DNA methylation in different regions on COAD remain unclear. This study analyzed the methylation of 12 different regions in COAD and normal colon tissue, revealing the differences in methylation distribution and the stability of methylation levels in COAD. Then, correlation analysis of differentially expressed genes and differentially methylated sites in different regions shows that methylation has an inhibitory effect on gene expression. Next, we performed Cox regression analysis and survival analysis based on clinical data, and identified three key genes (GSTM1, POU3F3, and RNF32) related to survival. We also constructed a prognostic risk scoring model, which demonstrated good predictive performance for prognosis. Furthermore, we identified three key DNA methylation CpG sites and their regulatory regions in three key genes. Subsequently, we constructed 120 COAD prediction models using 10 machine learning algorithms. The results show that these key genes and key DNA methylation CpG sites have very good predictive performance for COAD. Finally, we analyzed the immune microenvironment of three key genes and identified potential drugs associated with them. These results may provide theoretical support for the study of the regulatory mechanism of DNA methylation on gene expression, and may also provide potential biomarkers for the prognosis and early diagnosis of COAD.