Objective: To construct and verificate an RNA-binding protein (RBP)-associated prognostic model for gliomas using integrated bioinformatics analysis.
Methods: RNA-sequencing and clinic pathological data of glioma patients from The Cancer Genome Atlas (TCGA) database and the Chinese Glioma Genome Atlas database (CGGA) were downloaded. The aberrantly expressed RBPs were investigated between gliomas and normal samples in TCGA database. We then identified prognosis related hub genes and constructed a prognostic model. This model was further validated in the CGGA-693 and CGGA-325 cohorts.
Results: Totally 174 differently expressed genes-encoded RBPs were identified, containing 85 down-regulated and 89 up-regulated genes. We identified five genes-encoded RBPs (ERI1, RPS2, BRCA1, NXT1, and TRIM21) as prognosis related key genes and constructed a prognostic model. Overall survival (OS) analysis revealed that the patients in the high-risk subgroup based on the model were worse than those in the low-risk subgroup. The area under the receiver operator characteristic curve (AUC) of the prognostic model was 0.836 in the TCGA dataset and 0.708 in the CGGA-693 dataset, demonstrating a favorable prognostic model. Survival analyses of the five RBPs in the CGGA-325 cohort validated the findings. A nomogram was constructed based on the five genes and validated in the TCGA cohort, confirming a promising discriminating ability for gliomas.
Conclusion: The prognostic model of the five RBPs might serve as an independent prognostic algorithm for gliomas.
Keywords: RNA-binding protein; bioinformatics analysis; glioma; prognostic model.
© 2023. Huazhong University of Science and Technology.