Abstract: Ovarian cancer is a malignant tumor with low early diagnosis rate and high mortality rate.The identification of prognostic biomarkers and the prediction of patient risk are still important tasks in survival analysis. In this paper, a weighted gene co-expression network constructed by ovarian cancer prognosis-related genes was used to identify prognostic biomarkers and predict patient risk. Firstly, data of 320 patients with ovarian cancer were obtained from the TCGA(The cancer genome atlas) database, and 747 prognosis-related genes were selected by Cox univariable regression to construct a weight edgene co-expression network. Then, considering the biological significance of the network, the modulein co-expression network was re-weighted by integrating the protein-protein interaction(PPI)data from the module genes. Since topologically important genes in the network tend to play key roles in ovarian cancer, the topological properties of the genes in re-weighted network were used to rank the module genes. Finally, the Cox proportional hazards model was employed to construct prognostic models by these topologically important genes. By considering a balance between the model prediction ability and the number of genes, three prognostic biomarkers were identified. Survival analysis showed that the three biomarkers could significantly distinguish patients with different prognosis.

Key words: ovarian cancer, biomarker, prognostic model, weighted gene co-expression network analysis (WGCNA)