Identification And Comprehensive Analysis Of Relative Biomarkers In Hepatocellcular Carcinoma Based On Network And Pathways

Objective:Hepatocellular carcinoma?HCC?is a high mortality primary liver cancer.More than 700 thousand new HCC cases showed up every year in the world.It is one of the most common types of cancer in the past ten years,and it has the highest mortality rate in all kinds of cancer.Clinical diagnosis and treatment of HCC is hard because of its omplicated pathological mechanism.At the same time,the prognosis of hepatocellular carcinoma is also very bad,the prognosis of 2 years survival is less than 5%.Therefore,timely diagnosis and treatment is very important for the prognosis of HCC.Although many molecular biological studies have been devoted to exploring the underlying pathological mechanism of hepatocellular carcinoma in recent years,the molecular mechanism of HCC is still not fully understood.The purpose of this study is to collect and organize the relevant expression data,focus on the new biological targets in HCC through systematic biological analysis,discover the relationship between the biological processes and the various biological processes of HCC,search the disease subnetwork from the specific disease network based on the network analysis to get its related functions,and provide important support for subsequent research and clinical treatment.Method:1.Literature?Research?retrieval and admission criteria By retrieved specific microarray platform?GPL570?keyword in the Gene Expression Omnibus database?GEO,http://www.ncbi.nlm.nih.gov/geo/?,and filtered retrieval results which only contained hepatocellular carcinoma samples of patients with hepatocellular carcinoma or peripheral blood mononuclear cells?PBMC?as the candidate expression data set.2.Screening of differentially expressed genes By introduced in the Affy package on the R software platform,the original expression data of each candidate sample are read and preprocessed?background noise,standardization,etc.?.In addition,we utilized limma package to filtered expression data and got candidate differentially expressed genes?DEGs?,the threshold criterion was log?Fold Change??logFC?>1 or log?Fold Change??logFC?<-1and P.adjust<0.05?Benjamini&Hochberg?.A comprehensive gene matrix with the P value of each DEG in candidate sample set was given to get the DEGs with P_meta<0.01 during the Chi-square test.Then,combined disease related genes from the Genecards and the Phenolyzer database,the final comprehensive differential expression gene set was obtained.3.Enrichment analysis of HCC related DEGs Significant KEGG pathways were gained by introduced the over-representation analysis?ORA?method to up-regulated DEGs and down-regulated DEGs.The ClusterProfiler package were utilized to get significant enrichment results and figures of GO analysis.4.Crosstalk analysis of significant pathways Get the JC&OC scores of each pair of significant pathways?Pathway containing more than 5 genes?to get significant crosstalk results.Crosstalk results were visualized by CytoScape software in order to observing the interaction between pathways.5.Construction of human PPI network and HCC disease module The DEG set has been mapped according to the human protein interaction network?Protein-Protein Interaction Network,PPIN?,which contained 228096 protein interactions and 16022 proteins,in order to construct hepatocellular carcinoma disease specific network.Then utilized CytoScape software to get the visualization of the disease network,and MCODE plug-in to get HCC related disease modules.6.Comparison and analysis of HCC related network topology properties Summarize the node degree of the established specific network of hepatocellular carcinoma disease,compared with the node degree of the network constructed by the reference set of genes?which contained 595 cancer related genes from the cancer gene census database and 242 Parkinson disease related genes?.And estimated the topology characteristics of these network.Results:1.Literature?Research?retrieval and admission criteria Retrieving GEO database by keywords?GPL570?Affymetrix Human Genome U133 Plus 2.0 Array?[GEO Accession]?and?PBMC?[Title],we finally get 10 experiments datasets?by the end of May,2017?,contained 488 samples in total.2.Screening of differentially expressed genes After filtered gene expression set with threshold abs?log?Fold Change??>1 and P.adjust<0.05,we get the DEGs of each experiment dataset.By gathered all P values of DEGs from each dataset,we got the comprehensive candidate DEG matrix.Then used P_meta<0.01 during the Chi-square test to get comprehensive DEGs,we got 794 up-regulated and 959down-regulated DEGs.Finally,combined disease related genes from the Genecards and the Phenolyzer database,we got 444 DEGs as final HCC DEGs.3.KEGG enrichment analysis of HCC related DEGs In the ORA pathway enrichment procedure,we got 7 pathways from 177 up-regulated genes and 27pathways from 267 down-regulated genes?FDR<0.05?.These pathways were mainly related to p53 signaling pathway,chemical carcinogenesis,fatty acid degradation and arginine biosynthesis.In addition,GO enrichment result of Cluster-Profiler package has showed the top 10 GO terms from up-regulated and down-regulated DEGs.Up-regulated GO terms were mainly related to cell cycle,mitotic mitosis and chromatid separation in BP,while down-regulated GO terms were mainly contained response to substance or stimulus?inorganic substance,drug,metal ion and exrtacellular stimulus?and organic acid catabolic process in biological processes?BP?.5.Crosstalk analysis of significant pathways We finally got 171 crosstalk pairs from 39 significant pathways?in the enrichment analysis of whole DEGs?.After visualized in CytoScape,utilized MCODE to separated the crosstalk into two groups,one group was mainly about various pathways of cancer and virus infection?HTLV-1,Hepatitis B?,the other group was associated with chemical carcinogens,metabolism of lipid,development biology and energy metabolism.6.Construction of human PPI network and HCC disease module The final HCC disease specific network was contained 272 nodes and 528 edges.By visualization of CytoScape and finding disease modules with MCODE score>3,we finally got 5 HCC disease modules.KOBAS KEGG enrichment results showed the highest score module was related to steroid biosynthesis,chemical carcinogenesis,linoleic acid metabolism and tryptophan metabolism,the other modules were mainly associated with cell cycle,P53 signaling pathway,complement and coagulation cascade reaction.7.Comparison and analysis of HCC related network topology properties During the topology analysis of HCC related network,reference PD specific network and cancer related network,HCC specific network showed similar topology properties with cancer related network when the degree>200?Average degrees,HCC:42.3,CR:73.7,PD:32.9?.8.Survival analysis of hub genes We got the overlap genes of top 10 degree and top 10 betweeness in the HCC specific network,contained 6 hub genes:CDK1,MYC,CDKN1A,JUN,PCNA and SHC1.Among which CDK1 has the highest rank value?15.44?.In order to observed the relationship between expression and survival rate of these genes,we introduced survival analysis.Results showed the expression of JUN and CDK1 were significantly negatively correlated with survival rate?P<0.05?.Conclusions:1.The potential biological molecular mechanism of hepatocellular carcinoma is very complex.Data mining of this study showed its mainly related to viral infection,chemical carcinogenesis,developmental related biological pathways,P53 signaling pathways and abnormal metabolism of energy metabolism and biological pathways,which may lead to the occurrence and development of hepatocellular carcinoma.2.The central gene CDK1,MYC,CDKN1A,JUN,PCNA and SHC1 may play a key role in the occurrence and progression of hepatocellular carcinoma,especially the high expression of CDK1 and SHC1 has a significant negative correlation with the survival time.It can be used as a potential biological target for further further study of hepatocellular carcinoma.3.Compared with parkinson related gene sets,HCC related genes and cancer specific genes have higher connectivity,reflecting significant differences between cancer and psychiatric diseases.4.The method and analysis process used in this study can also serve as a research framework for other complex diseases,and provide some support for exploring more potential biological targets of complex diseases.