Identification of the subtypes of gastric cancer based on DNA methylation and the prediction of prognosis

Abstract

Background Gastric cancer (GC) is a digestive system cancer with a high mortality rate globally. Previous experiences and studies have provided clinicians with ample evidence to diagnose and treat patients with reasonable therapeutic options. However, there remains a need for sensitive biomarkers that can provide clues for early diagnosis and prognosis assessment. Results We found 610 independent prognosis-related 5′-cytosine-phosphate-guanine-3′ (CpG) sites (P < 0.05) among 21,121 sites in the training samples. We divided the GC samples into seven clusters based on the selected 610 sites. Cluster 6 had relatively higher methylation levels and high survival rates than the other six clusters. A prognostic risk model was constructed using the significantly altered CpG sites in cluster 6 (P < 0.05). This model could distinguish high-risk GC patients from low-risk groups efficiently with the area under the receiver operating characteristic curve of 0.92. Risk assessment showed that the high-risk patients had poorer prognosis than the low-risk patients. The methylation levels of the selected sites in the established model decreased as the risk scores increased. This model had been validated in testing group and its effectiveness was confirmed. Corresponding genes of the independent prognosis-associated CpGs were identified, they were enriched in several pathways such as pathways in cancer and gastric cancer. Among all of the genes, the transcript level of transforming growth factor β2 (TGFβ2) was changed in different tumor stages, T categories, grades, and patients’ survival states, and up-regulated in patients with GC compared with the normal. It was included in the pathways as pathways in cancer, hepatocellular carcinoma or gastric cancer. The methylation site located on the promoter of TGFβ2 was cg11976166. Conclusions This is the first study to separate GC into different molecular subtypes based on the CpG sites using a large number of samples. We constructed an effective prognosis risk model that can identify high-risk GC patients. The key CpGs sites or their corresponding genes such as TGFβ2 identified in this research can provide new clues that will enable gastroenterologists to make diagnosis or personalized prognosis assessments and better understand this disease.

Immunological Significance of Prognostic DNA Methylation Sites in Hepatocellular Carcinoma

“…It is, therefore, urgent to discover novel and promising factors that could predict response to ICB therapy for further individualized management and advanced precision immunotherapy ( Nishino et al, 2017 ; Mushtaq et al, 2018 ; Ng et al, 2020 ). Multiple studies demonstrated that DNA methylation might act as a pivotal player in prediction of response to therapy ( Yang et al, 2019 ; Li et al, 2020 ). In this study, we validated the DNA methylation–based risk score and potential hub targets (LRRC41 and KIAA1429) were significantly associated with expression level of ICB pivotal target genes (i.e., CTLA4).…”

Machine learning in the identification of prognostic DNA methylation biomarkers among patients with cancer: a systematic review of epigenome-wide studies

“…Based on selected CpGs in the prior stage, eight studies performed additional feature selection, using methods including LASSO (38, 45, 55, 57, 69, 78, 80), stepwise selection (38, 39), and the Akaike information criterion (64, 77), to further refine the number of CpGs. In contrast, four studies (23, 33, 34)(47) directly trained classification models with clusters as labels using Bayesian networks or random forest, while 17 studies constructed prognostic models (37-39, 45, 46, 53-57, 64, 69, 70, 75, 77, 78, 80). Lastly, models developed in the previous stage were mostly internally validated by split-sample validation, albeit four studies validated the model externally in an independent dataset (33, 50, 57, 78).…”

“…None of the studies using TCGA data reported median follow-up time, so little can be said about the data quality and interpretability of the conducted study. Second, six studies failed to specify the type of survival outcomes (e.g., overall survival or disease-free survival) (43, 46, 47, 60, 62, 75), which further limits the interpretation of results.…”

Differential DNA Methylation and Gene Expression Between ALV-J-Positive and ALV-J-Negative Chickens

“…In ovarian cancer, TGFB2 is overexpressed and plays a key role in ovarian oncogenesis by regulation of an epithelial-tomesenchymal transition (60). TGFB2 has been shown to be changed in different tumor stages, T categories, grades, and patients' survival states, and upregulated in patients with GC as compared with a normal control, and its expression could be affected by cg11976166 (61). In prostate cancer, a quantitative increase in promoter methylation levels of TGFβ2 are associated with PCa progression (62).…”