Establishment and validation of an RNA binding protein-associated prognostic model for ovarian cancer

Ovarian cancer (OC) is one of the most common gynecological malignant tumors worldwide, with high mortality and a poor prognosis. As the early symptoms of malignant ovarian tumors are not obvious, the cause of the disease is still unclear, and the patients' postoperative quality of life of decreases. Therefore, early diagnosis is a problem requiring an urgent solution.

We obtained the gene expression profiles of ovarian cancer and normal samples from TCGA and GTEx databases for differential expression analysis. From existing literature reports, we acquired the RNA-binding protein (RBP) list for the human species. Utilizing the online tool Starbase, we analyzed the interaction relationship between RBPs and their target genes and selected the modules of RBP target genes through Cytoscape. Finally, univariate and multivariate Cox regression analyses were used to determine the prognostic RBP signature.

We obtained 527 differentially expressed RBPs, which were involved in many important cellular events, such as RNA splicing, the cell cycle, and so on. We predicted several target genes of RBPs, constructed the interaction network of RBPs and their target genes, and obtained many modules from the Cytoscape analysis. Functional enrichment of RBP target genes also includes these important biological processes. Through Cox regression analysis, OC prognostic RBPs were identified and a 10-RBP model constructed. Further analysis showed that the model has high accuracy and sensitivity in predicting the 3/5-year survival rate.

Our study identified differentially expressed RBPs and their target genes in OC, and the results promote our understanding of the molecular mechanism of ovarian cancer. The current study could develop novel biomarkers for the diagnosis, treatment, and prognosis of OC and provide new ideas and prospects for future clinical research.