Screening Hub Genes Of Colon Cancer Based On TCGA Database And Establishment Of Prognostic Risk Model

Objective:Colon cancer is a highly malignant tumor of the gastrointestinal tract.Finding effective early diagnostic factors and new targets for drug therapy is the most important task in the diagnosis and treatment of colon cancer.The TCGA database covers colon cancer patients’clinical data,RNA-seq sequencing data,mi RNA sequencing data,etc.Therefore,in this study,we analyzed the relevant data of TCGA database and screened the hub genes of colon cancer using various bioinformatics methods,and also analyzed the factors related to the prognosis of colon cancer patients by constructing a prognostic analysis model of colon cancer.Method:Download the transcriptome sequencing data and the complete clinical data of colon cancer from TCGA database,the differentially expressed genes were screened by R software with corrected P<0.05,and the absolute value of differential expression multiple>4（FDR<0.05 and|log₂Fold Change|>2）as the condition.The 10hub genes with the highest correlation were screened by using the STRING database and Cyto Hubba plug-in,and the hub genes were further validated by the Oncomine database and quantitative real-time fluorescence PCR（q RT-PCR）.A prognostic analysis model for colon cancer patients was constructed,Cox regression analysis was performed,and model efficacy was assessed.Calculating the sample risk scores,divide the samples into two groups of high and low risk according to their medians,plot ROC curves of risk scores regarding survival time of 1,3 and 5 years for colon cancer patients,and calculate the corresponding AUC areas.Analyzing hub genes and clinical factors with K-M survival analysis to identify biomarkers and associated factors which affect prognosis.Results:1.In this study,we extracted 1665 differentially expressed genes of colon cancer from TCGA database,and identified the top 10 hub genes:CXCL1,GNG13,LPAR1,PPBP,CASR,NMUR2,PENK,CHRM2,SST,PYY.Being validated by Oncomine database,we obtained differentially expressed up-regulated hub genes CXCL1,PPBP,and down-regulated hub gene SST.2.The q RT-PCR results suggested that the m RNA expression levels of CXCL1and PPBP in colon cancer tissues were significantly higher than those in paracancerous tissues,and the differences were statistically significant,with P values both 0.004（P<0.05）.3.The results of single factor Cox regression analysis showed that age,clinical stage,CXCL1 and PPBP expression levels had prognostic value（P<0.2）.The Multifactorial Cox analysis results showed that clinical stage（stage III,stage IV）,CXCL1 and PPBP expression levels had significant prognostic value（P<0.05）.The prediction model C-index value=0.78,95%CI:0.743-0.813,ROC curve analysis yielded AUC values of 1,3 and 5 years survival time:0.74,0.73 and 0.69,respectively.So it means the model has good prediction accuracy and predictive ability.4.The results of K-M analysis suggested that age,clinical stage,risk score and CXCL1 expression level were statistically significant on the prognosis of colon cancer patients（P<0.05）.Conclusion:In this study,we successfully screened colon cancer hub genes with differential expression based on the TCGA database,and constructed a prognostic model with good predictive efficacy based on the hub genes and clinical factors of colon cancer.It has some significance for the diagnosis and prognosis of colon cancer patients.