Using the TCGA Database, Establishing and Validating an RNA-binding Protein-based on Prognostic Model for Liver Cancer

Tshetiz Dahal

Abstract

Background: The third most common cause of cancer-related death is liver cancer, one of the most prevalent hematological malignancies in the world. The term RNA-binding protein (RBP) refers to a group of proteins that bind to RNA to control metabolic processes. The development of RNA-binding proteins is related to the mortality of liver cancer victims. Methods: RBP gene expression data from liver cancer were extracted using the TCGA database. The differentially expressed RBPs were identified using enrichment analysis and volcano mapping (DE RBPs). Following that, the prognosis-related RBP genes were chosen using single-factor Cox regression analysis. The main prognosis-related RBPs were further selected using multi factor Cox regression analysis, yielding a formula for the patient's risk coefficient. Finally, a nomogram based on the patient's risk score was created and validated. Results: With the clinical information from each sample, we extracted 374 cancer tissue samples and 50 normal tissue samples. We screened 208 upregulated RBPs and 122 down regulated RBPs using enrichment analysis. EEF1E1, NOP56, UPF3B, SF3B4, SMG5, CD3EAP, BRCA1, BARD1, XPO5, CSTF2, EZH2, EXO1, RRP12, PRIM1, LIN28B, NROB1 and TCOF1 were high-risk genes for prognosis, while MRPL46, RCL1, MRPL54, CPEB3, IFIT5, PPARGC1A, EIF2AK On the prognosis-related RBPs, a multivariate Cox regression analysis was done, and the three major prognosis-related RBPs, BARD1, NR0B1, and EIF2AK4, were screened out. The patient risk coefficient was calculated using the formula shown below: risk rating = (1.207BARD1 Exp) + (0.483NR0B1 Exp) + (-0.720EIF2AK4 Exp). Finally, a nomogram based on the risk score was created to predict the survival time of patients from 1 to 5 years. Conclusion: The nomogram has good predictive value for the survival time of liver cancer patients.