Prognosis And Staging Research Of Hepatocellular Carcinoma Based On Weighted Gene Co-expression Network Analysis

Hepatocellular carcinoma(HCC)is the most common malignancy and the fifth leading cause of cancer-related death in the world.The development of HCC is a complex biological process with multiple genes,multifactor effects,and multistage progresses involved in.We know little about its development and progress mechanisms.At present,there are serious deficiencies in the determination of patients with high risk of recurrence and early diagnosis of hepatocellular carcinoma.Therefore,this article uses gene expression data to study its clinical manifestations at the intrinsic molecular level,and to understand the molecular mechanism of hepatocellular carcinoma and improve its diagnosis and treatment.First,this paper developed a prognostic model of HCC based on gene expression data.After preprocessing expression data and analysing differential expression genes(DEGs);we use a training set to cluster DEGs by weighted gene co-expression network(WGCNA)and clusters four modules.Then through the stepwise Cox risk proportional regression,we found the turquoise module is significantly correlated with patient survival,and the module gene is enriched in metabolic functions.Then we deeply explored this module by single factor Cox proportional hazard regression,so that we obtained 45 genes closely related to recurrence-free survival.The test set was used for verification.We examined the classification results by Kaplan-Meier curve and log-rank test,found that the recurrence rates of the low-risk marker group and the high-risk marker group were significantly different between the two groups.Finally,we combined the gene marker grouping and clinical information to verified that our discovery had a significant impact on HCC clinical.Secondly,in order to understand the relationship between the BCLC staging of HCC and gene expression,we establish a classification model for BCLC staging by WGCNA and random forest.Firstly,the data was pre-processed and differential expressed.The training set clustered five gene modules by WGCNA.Though enrichment analysis of gene modules,we found that the blue module genes are closely related to the development of HCC.Then PPI network analysis was carried out and visualized with Cytoscape.10 core genes with large connectivity in blue module were selected.Finally,these core genes were supervised studied by random forests,and applied to the test set.Our study found that the classification of early patients was greatly assisted,reaching 95.52%,but the classification effect of patients in the middle and late stages was not very satisfactory.We screened 45 genes from prognostic model and 10 core genes in BCLC staging studies.Some genes have been reported in related literature,indicating that screening genes do affect the occurrence and development of HCC at different degrees and aspects.Gene targets that do not understand the influence mechanism provide a forward-looking direction for our HCC study.The study in this paper has had a positive impact on the prognosis and staging of HCC.