Construction of a Mitochondrial Oxidative Stress-related Prognostic Risk Model for Hepatocellular Carcinoma Based on Machine Learning Algorithms
Liver cancer is one of the common malignant tumors,and the prognosis for advanced liver cancer is extremely poor.In view of the important role of mitochondrial oxidative stress in the development of hepa-tocellular carcinoma(HCC),mitochondrial oxidative stress-related genes were selected to construct a prog-nostic risk model for HCC.Firstly,the prognostic key genes were screened by using univariate Cox regres-sion analysis and three machine learning methods,namely support vector machine,random forest analysis,and LASSO regression analysis,and a model was constructed based on multivariate Cox regression.Secondly,the prognostic value of the model was further validated in the database.Thirdly,the possible mechanisms of the prognostic differences between the high-and low-risk groups were explored using gene enrichment analysis,and the immune microenvironment and treatment response between the two groups were compared.Finally,the expression of key genes in liver cancer tissues was verified by reverse transcription quantitative real-time PCR(RT-qPCR).Results showed that a total of 10 genes including PDE2A,TREM2,BMP6,NQO1,CPS1,EPO,MAPT,G6PD,SFN and HMOX1 were chosen out.Compared with the low-risk group,the high-risk group of HCC patients had a worse prognosis(P＜0.000 1).Enrichment analysis showed that peroxisome proliferator-activated receptor(PPAR)signaling pathway was significantly different between the high-and low-risk groups.And the tumor immunity analysis showed that the tumor immune infiltration,immune check-point-related genes,and immunotherapy response were also significantly different between the two groups.Validation results using RT-qPCR indicated that,compared with normal liver tissues,expressions of CPS1,PDE2A and BMP6 were lower in HCC tissues(P＜0.05),while expressions of G6PD and SFN were higher in HCC tissues(P＜0.05).In conclusion,the mitochondrial oxidative stress-related prognostic risk model estab-lished in this study has good predictive efficacy and accuracy,and can be used for the precise treatment of HCC.It would have a high clinical application value.
hepatocellular carcinoma(HCC)mitochondrial oxidative stressmachine learningprognosisrisk model
万方数据
维普
