Development And Validation of An RNA Binding Protein-Associated Prognostic Model for Colon Adenocarcinoma.

Abstract

Purpose: We aimed to identify prognostic RNA-binding proteins (RBP) in colon cancer, analyze their biological functions, and develop predictive models for patient prognosis. Materials and Methods: We downloaded COAD's RNA sequencing data from the Cancer Genome Atlas (TCGA) database, and the expression and prognostic value of these RBPs in COAD were systematically evaluated. Differential expression, KEGG, and GO enrichment analyses were then performed. Cytoscape was used to visualize the protein-protein interaction network, and Cox regression was used to establish a predictive model. Finally, the expression of RBP was verified by the HPA database and immunohistochemical staining. Results: A total of 472 differentially expressed RBPs were detected, including 321 up-regulated RBPs and 151 down-regulated RBPs. Four RBPs (MSI2, EZH2, NCL, TERT) were identified as key prognostic genes and used to construct prognostic models, based on this model, the overall survival (OS) of patients in high-risk subgroup was worse than that of patients in the low-risk subgroup. The area under the curve of time-dependent receiver operator characteristic curve of TCGA training set and Gene Expression Omnibus (GEO) validation set was 0.607 and 0.638 respectively, which confirmed that the prognosis model was good, it showed a good ability to identify COAD. Conclusion: In general, our prognostic model is based on 4 RBPs encoding genes, which greatly reduces the cost of sequencing and is more conducive to clinical applications.