A novel prognostic predictor for clear cell renal cell carcinoma constructed by weighted gene co-expression correlation network analysis

Abstract

Background Clear cell renal cell carcinoma (ccRCC) accounts the largest proportion of all types of renal tumor, it is highly invasive leading to shorter survival time of patients suffer from tumor. This study aims at constructing a new prognostic signature to provide a new reference for clinic work to adjust treatment protocols. Material and Methods Comprehensive gene expression profiles of ccRCC were downloaded from The Cancer Genome Atlas (TCGA). Weighted gene co-expression network analysis (WGCNA), a genome-wide profiling analysis, was applied to investigate key genes associated with clinical traits. Cox regression analysis was conducted to screen prognostic gene in ccRCC. LASSO analysis was used to enhance the robust of correlated genes. Then, a prognostic signature was developed as an independent factor to predict outcomes of ccRCC patients. Results The co-expression network was constructed by WGCNA, 26 modules were identified in total, including a module (grey) showed no significance. By associating the features with clinic features, we selected three modules (pink, purple and turquoise) significantly associated with the overall survival(OS) of ccRCC patients. Univariate Cox analysis for genes from each module separately and then we used LASSO regression to screen the significant genes for selecting potential prognosis element-genes. 8 genes (CAPRIN2, IFI44, LTV1, ZNF320, MTHFR, XPOT, BCL3 and PAX2) finally reserved by multivariate COX regression. ScoreW was defined as a prognosis modelestablished by the final potential prognosis 8 genes. Multivariate regression suggested the ScoreW was an independent prognosis predicting element. ROC curve method was performed and the AUC value of ScoreW showed satisfied and attractive clinical prognosis importance. Conclusion Our findings provided the framework of co-expression gene modules of ccRCC and identified valuable prognostic method might be detection for ccRCC patients.