DNA methylation patterns–based subtype distinction and identification of soft tissue sarcoma prognosis

Abstract

Soft tissue sarcomas (STSs) are heterogeneous at the clinical with a variable tendency of aggressive behavior. In this study, we constructed a specific DNA methylation-based classification to identify the distinct prognosis-subtypes of STSs based on the DNA methylation spectrum from the TCGA database. Eventually, samples were clustered into 4 subgroups, and their survival curves were distinct from each other. Meanwhile, the samples in each subgroup reflected differentially in several clinical features. Gene Ontology (GO) and the Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis was also conducted on the genes of the corresponding promoter regions of the above-described specific methylation sites, revealing that these genes were mainly concentrated in certain cancer-associated biological functions and pathways. In addition, we calculated the differences among clustered methylation sites and performed the specific methylation sites with LASSO algorithm. The selection operator algorithm was employed to derive a risk signature model, and a prognostic signature based on these methylation sites performed well for risk stratification in STSs patients. At last, a nomogram consisted of clinical features and risk score was developed for the survival prediction. This study declares that DNA methylation-based STSs subtype classification is highly relevant for future development of personalized therapy as it identifies the prediction value of patient prognosis.

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

“…Among the 39 studies that adopted clustering-based workflows (5, 18-23, 29, 32-35, 37-39, 43, 45-47, 50, 53-57, 61, 62, 64, 69, 70, 75, 77-80, 83, 84, 87, 88), most started with a preliminary screening of CpG sites (Figure 4.1). The most frequently used pre-selection approach was selecting CpG sites that were significantly associated with prognosis by first performing univariable Cox regression analysis, and then performing multivariable Cox analysis with adjustment for clinical variables.…”

“…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).…”

“…As a remedy semi-supervised approaches might be used, which could combine the best of both approaches. Alternatively, studies might use unsupervised and supervised approaches in conjunction at different stages to select features, as done in some of the studies (38, 45, 55, 57, 69, 78, 80).…”

“…Lastly, we found that 18 identified studies first constructed a Cox prognostic model based on single selected CpG sites (i.e., methylation score), which was subsequently used as a single predictor, together with other clinical variables, to construct a second Cox prognostic model (6, 31, 36, 39-42, 55, 57, 66-69, 73, 78, 80, 81). It might be an inappropriate approach to construct the first “methylation score”, in which the relative weight of individual CpGs were consequently fixed for the second prognostic model.…”

Can DNA Methylation Profiling Classify Histologic Subtypes and Grades in Soft Tissue Sarcoma?

“…Human biospecimen and associated data used in this study were provided by the St. Vincent’s Hospital Biobank of The Catholic University of Korea (47SA2022002-001 and 47SA2022003-001). Genomic annotation of DNA methylation sites and hierarchical cluster analysis were performed to interpret DNA methylation profiling results, according to a previous study [11]. All statistical analyses were conducted using R version 4.1.1 (http://www.r-project.org).…”