A prognostic signature based on three non‐coding RNAs for prediction of the overall survival of glioma patients. Issue 4 (7th March 2019)

• Record Type: Journal Article
• Authors: Xian, Junmin; Zhang, Quanzhong; Guo, Xiwen; Liang, Xiankun; Liu, Xinhua; Feng, Yugong
• Is Part Of: FEBS open bio. Volume 9:Issue 4(2019)
• Page Start: 682
• Page End: 692
• Publication Date: 2019-03-07
• Subjects: GEO -- glioma -- GSEA -- non‐coding RNA -- prognostic signature -- random survival forest; Molecular biology -- Periodicals; Cytology -- Periodicals; Life sciences -- Periodicals; Biological Science Disciplines -- Periodicals; Molecular Biology -- Periodicals; Cell Biology -- Periodicals; Cytology; Life sciences; Molecular biology; Periodicals; 572.805
• DOI: 10.1002/2211-5463.12602
• Languages: English
• ISSNs: 2211-5463

Abstract

Abstract : Recent studies have identified certain non‐coding RNAs (ncRNAs) as biomarkers of disease progression. Glioma is the most common primary intracranial cancer, with high mortality. Here, we developed a prognostic signature for prediction of overall survival (OS) of glioma patients by analyzing ncRNA expression profiles. We downloaded gene expression profiles of glioma patients along with their clinical information from the Gene Expression Omnibus and extracted ncRNA expression profiles via a microarray annotation file. Correlations between ncRNAs and glioma patients' OS were first evaluated through univariate Cox regression analysis and a permutation test, followed by random survival forest analysis for further screening of valuable ncRNA signatures. Prognostic signatures could be established as a risk score formula by including ncRNA signature expression values weighted by their estimated regression coefficients. Patients could be divided into high risk and low risk subgroups by using the median risk score as cutoff. As a result, glioma patients with a high risk score tended to have shorter OS than those with low risk scores, which was confirmed by analyzing another set of glioma patients in an independent dataset. Additionally, gene set enrichment analysis showed significant enrichment of cancer development‐related biological processes and pathways. Our study may provide further insights into the evaluation of glioma patients' prognosis. Abstract : Through integrating univariate and multivariate Cox regression analysis of a glioma expression dataset from the Gene Expression Omnibus, we developed a prognostic model that contains three non‐coding RNAs, through which a risk score was allocated to each sample. Glioma patients with a high risk score tended to have shorter overall survival than those with a low risk score, which was confirmed using another set of glioma patients in an independent dataset.