Extended Data Fig. 3: Glioma Methylation Signature validation.
a, Histogram depicting the distribution of GMS-scores of CATNON DNA-methylation samples. Bars are colored based on the assigned GMS-class (‘hyper’: >1 M; ‘intermediate’: 0 – 1 M; ‘hypo’: <0 M). b, Kaplan-Meier curve for OS when stratifying patients of the TCGA cohort on the GMS-class. The hypo GMS-class was negatively associated with OS (‘hypo’, P < 0.001, HR = 4.0 (2.21-7.39); ‘intermediate’, P = 0.21, HR = 1.4 (0.81-2.51)). c-d, Kaplan-Meier curves for OS when stratifying based on the GMS-class of WHO grade 2 and grade 3 patients of the TCGA cohort, respectively. As compared to the hyper GMS-class, the hypo GMS-class was negatively associated with OS in WHO grade 2 tumors (c) (‘hypo’, P = 0.002, HR = 3.7495 (1.6344-8.601); ‘intermediate’, P = 0.316, HR = 1.4868 (0.6843-3.230)), and WHO grade 3 tumors (d) (‘hypo’, P = 0.004, HR = 3.9954 (1.55631-10.257); ‘intermediate’, P = 0.765, HR = 1.1407 (0.4812-2.704), two-sided Wald test). e, Output of multivariate CoxPH model on OS using data of the TCGA cohort, with the hyper GMS-class and WHO grade 2 as reference. f, Histogram depicting the distribution of GMS-scores of 5-hmc DNA-methylation samples of Glowaka et al.20. Cut-point for high and low GMS-scores was set at the median (red dotted line). g, Volcanoplot depicting the differential hydroxy-methylation analysis between GMS high (n = 21) and low (n = 21) samples of Glowaka et al. None exceeded the significance threshold of <0.05.
