Identification Of Hub Genes In Cholangiocarcinoma Using Weighted Gene Co-expression Network Analysis

Objective:Cholangiocarcinoma is a malignant tumor of the digestive system originating from the epithelial cells of the bile duct mucosa.Patients with early cholangiocarcinoma often have no clinical symptoms.As the disease progresses,symptoms such as abdominal discomfort,fatigue,nausea,jaundice and fever may appear.Because of the occult onset of cholangiocarcinoma,most patients have no symptoms or no specific symptoms in the early stage,which poses a great challenge to the early diagnosis of cholangiocarcinoma.In addition,many patients have little understanding of the disease and poor consciousness,which leads to a low rate of early detection of patients with cholangiocarcinoma.Most of them are in the advanced stage after detection,so they lose the opportunity for surgical treatment and can only adopt some conservative treatment methods,which seriously affects the prognosis and quality of life of patients.Therefore,the aim of this study was to find the pivotal gene of cholangiocarcinoma patients so as to guide the screening of high-risk population and the targeted treatment of cholangiocarcinoma patients.Methods:In this study,gene expression data(including RNA-seq count,FPKM)and clinical data of cholangiocarcinoma were downloaded by searching the cancer genome database(TCGA),including 33 cancer tissue samples and 8 para-cancer tissue samples.We divided the gene samples into tumor tissue and para-carcinoma tissue samples,and screened the gene expression data(RNA-seq count)for differentially expressed genes(DEGs)through the edge R package to remove the low-expressed and non-differentially expressed genes.The downloaded RNA-seq FPKM list was compared with the DEGs obtained in the previous step and the intersection was taken to obtain the FPKM of the DEGs for subsequent data analysis.Since 6 samples in the clinical data were missing in TNM grade data,we first removed the samples containing missing values.Weighted Gene co-expression Network Analysis(WGCNA)are greatly influenced by outlier samples,so we need to perform cluster analysis to remove outlier sample data.WGCNA was used to construct the gene co-expression network,divide the gene modules and find the genes closely related to the tumor classification.Gene Ontology(GO)and KEGG pathway were analyzed for the modules and genes closely related to tumor classification.Results:WGCNA of DGEs identified 14 modules.combining the module with clinical traits,we found that the correlation coefficient between the red module and the TNM grade of cancer was 0.43,P=0.01,which was the most closely related module to the TNM grade of cancer.Screening hub gene of the red module,we obtained hub genes such as FGF2,FGFBP1,EPHA5,CELF3,NKX6-1,CTRC,CPB1 and TAGLN3.searching the STRING database,we found that FGFBP1 is closely related to the fibroblast growth factor family(FGFs),which plays a biological role by binding unactivated FGFs to the extracellular matrix,dissociating it from the extracellular matrix and presenting it to FGFs receptor after activation.At present,the targeted drugs acting on FGFR2 are one of the research hotspots in the treatment of cholangiocarcinoma.We imported the red module genes into the Web Gestalt online analysis tool for enrichment analysis,The genes of this module were mainly enriched in epithelial cell proliferation,regulation of epithelial cell proliferation,positive regulation of epithelial cell proliferation,animal organ morphogenesis,positive regulation of cell proliferation,cell proliferation,glycosaminoglycan binding,tube development,endothelial cell proliferation,signal release,etc.Their function is to participate in the activation or increase of epithelial cell proliferation,and the process of signal secretion from cell source or release to extracellular matrix.Since other modules were also significantly correlated with tumor classification,the network Screening function was used to screen genes related to this clinical trait.We set the q.weighted threshold to 0.05 and obtained 196 genes that were highly correlated with tumor grade.Using the STRING online tool to analyze the interlacing network of egg albumen with 196 related genes,we found that the genes with high connectivity were APOB,KNG1,AHSG,HRG,FGB,SPARC,etc.Then,we imported these genes into the Web Gestalt online analysis tool for enrichment analysis,Related genes were enriched in extracellular structure organization,collagen-containing extracellular matrix,extracellular matrix,small molecule catabolic process,extracellular matrix organization,extracellular matrix structural constituent conferring tensile strength,Protein digestion and absorption,Glycolysis /Gluconeogenesis,Fatty acid degradation,Metabolic pathways,etc.It is suggested that the increase of tumor invasiveness is closely related to the structure of extracellular matrix.Conclusions:(1)According to the results of WGCNA,FGF2,FGFBP1,EPHA5,CELF3,NKX6-1,CTRC,CPB1,TAGLN3 may be hub genes for the occurrence,invasion and metastasis of cholangiocarcinoma.At the same time,PPI network analysis results suggested that APOB,KNG1,AHSG, HRG,FGB,SPARC may be strongly correlated with the occurrence,invasion and metastasis of cholangiocarcinoma.At present,relevant studies have confirmed that some of the above genes play an important role in the occurrence and metastasis of certain tumors and described the mechanism of action,but whether these genes have similar roles in the occurrence,invasion and metastasis of cholangiocarcinoma still needs to be verified.(2)These genes are likely to be used to detect early cholangiocarcinoma,targeted therapy and predict the prognosis of patients with cholangiocarcinoma.