Identification and validation of a five-lncRNA signature for predicting survival with targeted drug candidates in ovarian cancer
Nuan Lin, Jia‐zhe Lin, Yoshiaki Tanaka, Pingnan Sun, Xiaoling Zhou
Abstract
The dysregulation of long non-coding RNAs (lncRNAs) plays a crucial role in ovarian cancer (OC). In this study, we screened out five differentially expressed lncRNAs (AC092718.4, AC138035.1, BMPR1B-DT, RNF157-AS1, and TPT1-AS1) between OC and normal ovarian based on TCGA and GTEx RNA-seq databases by using Kaplan–Meier analysis and univariate Cox, LASSO, and multivariate Cox regression. Then, a risk signature was constructed, with 1, 3, 5-year survival prediction accuracy confirmed by ROC curves, and an online survival calculator for easier clinical use. With lncRNA-microRNA-mRNA regulatory networks established, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed, suggesting the involvement of a variety of cancer-related functions and pathways. Finally, five candidate small-molecule drugs (thioridazine, trifluoperazine, loperamide, LY294002, and puromycin) were predicted by Connectivity Map. In conclusion, we identified a 5-lncRNA signature of prognostic value with its ceRNA networks, and five candidate drugs against OC.