基于WGCNA方法的肝癌预后研究

【Objective】Hepatocellularcarcinoma (HCC) is an aggressive malignancy. At present, it is challenging to identify patients with high risk of recurrence, which would warrant additional therapies. By applying weighted gene co-expression network analysis (WGCNA) and Cox proportional hazards regression analysis, we developed a gene-expression model for identifying genes that may serve as markers for HCC prognosis.【Methods】In this paper, the weighted gene co-expression network is used to cluster the differentially expressed genes in the training set to obtain four modules.Then using multivariate Cox proportional hazards regression analysis with stepwise selection, we select the modules that were significantly associated with patient survival time and perform functional enrichment and pathway enrichment analysis to understand the reasons for clustering. Finally,univariate cox proportional hazards regression analysis model was established to screen out the closely related genes in the module to evaluate the relationship between gene expression level and non-recurrence rate. The test set is used for verification. Then draft Kaplan-Meier curves. The rank test was performed, and the gene marker grouping and clinical information were used as covariates for multivariate cox regression analysis.【Results】We obtained 45 genes from the turquoise module that are closely related to recurrence-free survival. The genes in this module are closely related to the metabolic processes of cells.Not only variables such as BCLC staging are associated with the prognosis of hepatocellular carcinoma, but these gene markers can also be used as an indicator of the prognosis of hepatocellular carcinoma. The findings of this study can improve the understanding of the molecular pathogenesis and prognosis of hepatocellular carcinoma.