Development and Validation of Epigenetic Modification-Related Signals for the Diagnosis and Prognosis of Hepatocellular Carcinoma.

Background: Increasing evidence has indicated that abnormal epigenetic factors such as RNA m6A modification, histone modification, DNA methylation, RNA binding proteins and transcription factors are correlated with hepatocarcinogenesis. However, it is unknown how epigenetic modification-associated genes contribute to the occurrence and clinical outcome of hepatocellular carcinoma (HCC). Thus, we constructed the epigenetic modification-associated models that may enhance the diagnosis and prognosis of HCC.

Methods: In this study, we focused on the clinical value of epigenetic modification-associated genes for HCC. Our gene expression data were collected from TCGA and HCC data sets from the GEO database to ensure the reliability of the data. Their functions were analyzed by bioinformatics methods. We used lasso regression, Support vector machine (SVM), logistic regression and Cox regression to construct the diagnostic and prognostic models. We also constructed a nomogram of the practicability of the above-mentioned prognostic model. The above results were verified in an independent liver cancer data set from the ICGC database and clinical samples. Furthermore, we carried out pan-cancer analysis to verify the specificity of the above model and screened a wide range of drug candidates.

Results: Many epigenetic modification-associated genes were significantly different in HCC and normal liver tissues. The gene signatures showed a good ability to predict the occurrence and survival of HCC patients, as verified by DCA and ROC curve analysis.

Conclusion: Gene signatures based on epigenetic modification-associated genes can be used to identify the occurrence and prognosis of liver cancer.