ABSTRACT The m 5 C RNA modifications have been implicated in the pathogenesis of urothelial carcinoma and hold potential as prognostic biomarkers for muscle‐invasive bladder cancer (MIBC) patients. In this study, we developed an MIBC‐risk model by integrating m 5 C modification‐related genes and differentially expressed genes using Nanopore sequencing and a machine learning approach. Compared to our previous research, we observed that m 5 C modifications are more functional, with the most enriched regions being the 3′UTR and exons. Our analysis revealed differential m 5 C methylation sites in several well‐characterized cancer‐related genes, including BMI1, PTEN, MALAT1, FADD, STAT5A, BIRC6, FOXO3, CCNG1, PAK2, UBE2L3, SMARCB1, and TUG1. Functional enrichment analysis demonstrated significant involvement of these genes in key oncogenic pathways, particularly DNA damage response, double‐strand break repair, p53 signaling, MAPK cascade, NF‐κB signaling, and cell proliferation/migration pathways. Unlike models based on single factors, the combination of m 5 C modification‐related genes and differentially expressed genes resulted in a more effective classification model. This approach yielded an optimized 11‐gene prognostic signature comprising GGA1, NUMBL, ECHDC2, NLRC5, EIF2D, GJA1, XPC, DAZAP2, C6orf120, WDR45, and CES1, which demonstrated superior predictive performance in TCGA MIBC patients. These findings establish m 5 C RNA modification patterns as promising molecular signatures for MIBC prognosis and potential therapeutic targets.
Zhang et al. (Thu,) studied this question.