Identification of a RNA-Seq based prognostic signature with five lncRNAs for lung squamous cell carcinoma

Abstract

Long non-coding RNAs (lncRNAs) expression profile signature for survival assessment in lung squamous cell carcinoma (LUSC) are largely inconsistent due to distinct detecting approaches and small sample size. Systematic and integrative investigation of RNA-Seq based data from The Cancer Genome Atlas (TCGA) herein was performed to determine candidate lncRNAs for prognosis evaluation of LUSC. A total of 60483 genes, including 7589 lncRNAs were assessed in a cohort including 478 LUSC cases with follow-up data. Firstly, 4225 differentially expressed lncRNAs were obtained via R packages. Next, univariate and multivariate Cox proportional hazards regression revealed that 41 lncRNAs were closely related to the survival of LUSC. Finally, lncRNA based prognosis index (PI) could predict overall survival of LUSC with high accuracy (AUC = 0.652, CI: 0.598, 0.705), PI = exp_CYP4F26P*β_CYP4F26P+exp_{RP11-108M12.3}*β_{RP11-108M12.3}+exp_RP11-38M8.1*β_RP11-38M8.1+exp_RP11-54H7.4*β_RP11-54H7.4+exp_ZNF503-AS1*β_ZNF503-AS1. Furthermore, it was confirmed that the five-lncRNA signature could act as an independent prognostic indicator for LUSC (HR = 2.068, p < 0.001 with univariate analysis, HR = 1.928, p = 0.038 with multivariate). Besides, we constructed a weighted gene co-expression network analysis (WGCNA) of key lncRNA RP11-54H7.4 according to the p-value of related genes' weight. This study provides a RNA-Seq based prognostic signature with five lncRNAs for further clinical application to LUSC patients.