Systematic profiling of alternative splicing signature reveals prognostic predictor for ovarian cancer

Objective. The majority of genes are alternatively spliced and growing evidence suggests that alternative splicing is modified in cancer and is associated with cancer progression. Systematic analysis of alternative splicing signature in ovarian cancer is lacking and greatly needed. Methods. We profiled genome-wide alternative splicing events in 408 ovarian serous cystadenocarcinoma (OV) patients in TCGA. Seven types of alternative splicing events were curated and prognostic analyses were performed with predictive models and splicing network built for OV patients. Results. Among 48,049 mRNA splicing events in 10,582 genes, we detected 2,611 alternative splicing events in 2,036 genes which were significant associated with overall survival of OV patients. Exon skip events were the most powerful prognostic factors among the seven types. The area under the curve of the receiver-operator characteristic curve for prognostic predictor, which was built with top significant alternative splicing events, was 0.937 at 2,000 days of overall survival, indicating powerful efficiency in distinguishing patient outcome. Interestingly, splicing correlation network suggested obvious trends in the role of splicing factors in OV. Conclusions. In summary, we built powerful prognostic predictors for OV patients and uncovered interesting splicing networks which could be underlying mechanisms. (C) 2017 Elsevier Inc. All rights reserved.