Screening Of Differentially Expressed Genes In Gastric Cancer Based On Bio-information Analysis And Establishment Of Prognostic Risk Model

Background:In recent years,with the continuous development of science and technology and the continuous improvement of medical technology,genomics has been rapidly developed in China and widely applied in clinical practice.In the context of big data,deep mining of tumor bioinformation has enabled us to have a better understanding of cancer.The screening of differentially expressed genes(DEGs)based on bioassay has been applied to search for potential biomarkers in gastric cancer.However,due to the lack of clinical samples at present,the genes studied may not be accurate.Therefore,it is urgent to explore effective molecular markers of gastric cancer.Objective:Based on bioinformatics analysis,we explored and screened DEGs with clinical value in gastric cancer and constructed an effective prognostic risk assessment model.Methods:This study was mainly carried out from two aspects.First,7 groups of gastric cancer gene data sets were retrieved from the world authoritative database GEO,and The "LIMMA" package and robust rank aggregation(RRA)method in The R Programming Language were used to screen DEGs from The 7 sets of data sets successively for functional annotation and enrichment analysis.Secondly,Hub genes were screened by a variety of bioinformatics analysis methods,and The Cancer Genome Atlas(TCGA)was used to verify The differential expression results of The above key genes.Proportional risk regression(Cox)analysis was used to establish a prognostic risk model.Finally,the prognostic value of the above models was evaluated using GSE84433 in GEO.Results:According to the results obtained in the first part of this study,288 DEGs were screened by the RRA algorithm,of which 135 were significantly up-regulated and 153 were significantly down-regulated.The up-regulated genes were significantly related to biological processes such as cell adhesion,and were enriched in signaling pathways such as P53.The down-regulated genes were significantly involved in energy metabolism and binding.They were also found to be enriched in various metabolic pathways.In the second part,11 key genes such as SPARC were screened out based on the results of the first part,and the enrichment analysis showed that the above genes were significantly enriched in PI3K-Akt and other signaling pathways.Subsequently,a prognosis evaluation model of gastric cancer based on 7 genes was established by using multi-factor Cox proportional risk regression model and TCGA expression profile and patient prognosis and survival information.Receiver operating characteristic(ROC)curve analysis results showed that this model had a high accuracy in predicting the prognosis of gastric cancer patients(AUC = 0.802).Kaplan-Meier(K-M)curve diagram suggested a significantly higher overall survival rate in the low-risk group compared with the high-risk group(log-rank test P-value <0.001).The 5-year survival rates in the high-risk and low-risk groups were 10.9percent and 55.8 percent,respectively.Finally,GSE84433 in the GEO database was used to verify the good prognostic value of this model.Conclusion:(1)Based on bioinformatics analysis,288 significant DEGs were screened in gastric cancer,of which 135 were significantly up-regulated and 153 were significantly down-regulated.(2)The up-regulated genes were significantly involved in biological processes such as cell adhesion,adhesion and cytoskeleton activity,while the down-regulated genes were involved in metabolism of various substances and glycolysis pathways.(3)According to the DEGs screened earlier,11 hub genes were obtained by HIPPIE database combined with Fisher’s accurate test.These genes were significantly enriched in ECM receptor interaction,adhesion plaque,cancer proteoglycan and other pathways.(4)Eleven key genes were analyzed by multivariate Cox risk regression model,and finally a prognosis evaluation model based on 7 DEGs of gastric cancer was constructed.The results of ROC analysis showed that this model had a high accuracy in predicting the 5-year survival rate of gastric cancer patients(AUC = 0.802).Overall survival(OS)of high-risk and low-risk patients was compared using the Kaplan-Meier method based on the median risk score.5-year survival rates in the high-risk and low-risk groups were 10.9% and 55.8%,respectively,and the results were statistically significant.(5)External data set GSE84433 was used to prove that this model has good prognostic value.