Analysis Of Gene And MicroRNAs Expression Profile And Clinical Correlation Of Prostate Cacinoma In TCGA Database By Bioinformatics

Objective:The purpose of this study was to use prostate cancer RNA-Seq data provided in TCGA to screen potential prostate cancer related genes and miRNA.Screening of key genes related to Prostate Cancer by establishing protein interaction Network.The regulatory network of miRNA and target genes was established to screen prostate cancer related miRNA,so as to expand the knowledge base of prostate cancer molecular biology and provide direction for functional research and targeted therapy.Finally,in order to find the molecular markers to evaluate the prognosis of prostate cancer,guide risk stratification and formulate treatment strategies,the clinical correlation between differentially expressed miRNA and genes was explored.Methods:The RNA-SEQ data and clinical data of mRNA and miRNA from 499 prostate cancer tissue samples and 52 adjacent normal prostate tissue samples were downloaded from TCGA database.The DESeq was used to analyse the RNA-seq data.DEGs and differentially expressed miRNA were screened respectively.Functional annotations of DEGs were conducted by gene ontology(GO)and Kyoto encyclopedia of genes and genomes(KEGG).Then the protein interaction network(PPI),was constructed to screen hub genes by using Cytoscape.For differentially expressed miRNA,miRwalk2.0 Target Gene Prediction website was used to analyse which combine 12 authoritative miRNA target gene prediction sites to screen the target gene.Then overlapping the target genes with differentially expressed genes involved in tumor signal regulation to screen differentially expressed miRNA involved in tumor signaling pathway and its target genes.Finally the miRNA data set in GEO database is used to verify the differential expression of miRNA.The clinical correlation between DEGs and differentially expressed miRNA was explored by analyzing the clinical data in TCGA.Cox univariate regression and Cox multivariate regression was used to establish a risk assessment model.Results:1.A total of 2347 DEGs were obtained,of which 712 were up-regulated and 1635 were down-regulated.The biological enrichment function of DAVID database showed that DEGs was involved in tumor signal transduction.Seven hub genes were screened by analyzing PPI.2.A total of 50 differentially expressed miRNA were obtained,of which 23 were up-regulated and 27 were down-regulated.By using miRwalk2.0 online prediction website,14 miRNA and 25 target genes were obtained which involved in prostate cancer signal regulation.Finally,the differential expression of 10 miRNA is verified by analyzing the data of 14 data sets in GEO database.3.A risk assessment model composed of FGF7,LPAR1,miR-3664,miR-5680,miR-5687 was established.The model has high sensitivity and specificity in predicting the survival risk of prostate cancer patients,which can be used as a reliable biomarker to predict the prognosis of prostate patients.4.Four miRNAs and two DEGs was found that related to the invasion characteristics of prostate cancer,which have the potential to predict the risk of prostate cancer.Conclusion:Based on the large biological information data of TCGA database,we analyzed the expression profile of miRNA and genes related to prostate cancer.Sevent hub genes were screened by conducting PPI.The network of miRNA with target gene was drawn to screen the miRNA,which related to growth and development of prostate cancer,The above research results provide direction for follow-up functional research.A miRNA-gene model was established to predict the prognosis of patients with prostate cancer,some differentially expressed miRNA and DEGs show clinical correlation,which can guide the prognosis evaluation and risk stratification of prostate cancer in clinical practice.