Identification of molecular biomarkers for ovarian cancer using computational approaches.

Ovarian cancer is one of the major causes of mortality among women. This is partly because of highly asymptomatic nature, lack of reliable screening techniques and non-availability of effective biomarkers of ovarian cancer. The recent availability of high-throughput data and consequently the development of network medicine approach may play a key role in deciphering the underlying global mechanism involved in a complex disease. This novel approach in medicine will pave the way in translating the new molecular insights into an effective drug therapy applying better diagnostic, prognostic and predictive tests for a complex disease. In this study, we performed reconstruction of gene co-expression networks with a query-based method in healthy and different stages of ovarian cancer to identify new potential biomarkers from the reported biomarker genes. We proposed 17 genes as new potential biomarkers for ovarian cancer that can effectively classify a disease sample from a healthy sample. Most of the predicted genes are found to be differentially expressed between healthy and diseased states. Moreover, the survival analysis showed that these genes have a significantly higher effect on the overall survival rate of the patient than the established biomarkers. The comparative analyses of the co-expression networks across healthy and different stages of ovarian cancer have provided valuable insights into the dynamic nature of ovarian cancer.