Fig. 1

Organization of DNA methylation by dataset and GBM subtype. The representation of the twenty first eigenvalues shows the organization of the inertia structure a given by principal component analysis (PCA) of the DNA methylation data. The first vectorial plane, defined by the two first eigenvalues, represents 21% of the table inertia. The patient samples are represented on the first vectorial plane of the PCA (axes 1 and 2) of DNA methylation, annotated by study origin (DKFZ, red; EORTC & LN-Pilot, green; Nordic, blue; TCGA, pink; due to the overlap of the datasets, only one label is visible) b and methylation-based GBM classification c, MES, red; RTK I, green; and RTK II, blue. The R-squared (R²) and p-value (99 permutations, ADONIS) testing the effect between the DNA methylation data and dataset b and GBM subtype c are indicated. The boxplot representations for the patients age d, sample purity e, and DNAm age acceleration f are drawn, stratified by methylation-based GBM classification and study origin. No age related association with the three GBM subtypes was observed (p = 0.806, Wald’s test) d, while significant differences were found for HM-purity (p < 0.001, Wald’s test) e, and for DNAm age acceleration (p < 0.001, Wald’s test) f