Extended Data Fig. 6: DNA methylation changes upon AZA treatment in human AML patients.
a, WGBS was performed from human AML samples which either carried a DNMT3AR882H mutation (DNMT3A-R882H, n = 1) or which were wildtype for DNMT3A (WT, n = 2). For each patient a sample was available at diagnosis (Control) and at day 18 of treatment with AZA. Pairwise DMRs were called across genotypes and timepoints and categorized as having high or low levels of DNA methylation. DMRs from human AML samples were tested for enrichment of murine DMR clusters as described in Figure 5. Depicted are log10(qvalues)*sign(log_odds_ratio) from dark-blue (depleted features) to red (enriched features) indicating the strength of enrichment of a given murine DMR cluster as compared to all other clusters. Only DMR categories with significant enrichments are shown. b, DMRs from human AML samples were tested for enrichment of MSigDB hallmark gene sets. Depicted are –log10(q-values)*sign(log_odds_ratio) from dark-blue (depleted features) to red (enriched features) indicating the strength of enrichment of a given feature in a particular cluster as compared to all other clusters. Only DMR categories with significant enrichments are shown. c, ERV loci shown to be affected by AZA in human colorectal cancer cells6. IGV browser tracks display DNA methylation data for MER4A1 locus, MER50 locus, MER57 locus and MLT1C locus. Each vertical line represents a CpG dinucleotide and the height of the line represents the DNA methylation in percent.
