Identification and validation of RNA-binding protein-associated biomarkers in esophageal cancer

Abstract

Esophageal cancer (ESCA) is highly prevalent globally, but the role of RNA-binding proteins (RBPs) in it remains unclear. ESCA datasets were analyzed to identify differentially expressed genes (DEGs). Candidates were derived by intersecting DEGs with RBP-related genes and validated via machine learning, expression analysis, and receiver operating characteristic (ROC) curves. Prognostic value, functional enrichment, immune microenvironment, regulatory networks, drug prediction, and Eca-109 cell validation were assessed. Four biomarkers (CDC20, COL7A1, DNMT3B, UBE2T) were confirmed. Low UBE2T/DNMT3B correlated with better survival. In addition, CDC20/UBE2T were linked to DNA replication; COL7A1/DNMT3B to olfactory transduction. Immune analysis showed 15 cell types differed significantly, with most chemokines negatively correlating with biomarkers. COL7A1, DNMT3B, and UBE2T were associated with transcription factor TET1. The four biomarkers corresponded to 79, 26, 67, and 23 targeted drugs, respectively, and all were highly expressed in Eca-109 cells, advancing insights into RBP-related mechanisms in ESCA.