Extended Data Fig. 4: TWGBS reveals preferential DNA hypomethylation of endogenous retroviral elements and interferon response genes in AZA treated DNMT3AR882H mutant mice.
a, Correlation analysis of DNA methylation levels in all 500 bp tiles using Pearson correlation comparing NaCl-treated +/+ with +/m LSK cells. PCC *100 = Pearson Correlation Coefficient multiplied by 100. b, Principal component analysis based on DNA methylation levels in genome-wide 500 bp tiling windows. red: +/+ NaCl control (n = 5 mice), blue: +/+ AZA (n = 5 mice), green: +/m NaCl control (n = 3 mice), purple: +/m AZA (n = 2 mice). c, Heatmap showing hierarchical clustering of mean beta values (i.e. % DNA methylation) for all 15,468 DMRs across all independent experiments (mice) analyzed. For mouse numbers per genotype and treatment see panel b. d and e, IGV browser tracks displaying DNA methylation data for all independent experiments analyzed by TWGBS in the present study. For mouse numbers per genotype and treatment see panel b. Each vertical line represents a CpG dinucleotide and the height of the line represents the DNA methylation in percent. Depicted is the (d) Ikzf1 gene locus which features 6 DMRs and the (e) Irf8 gene locus which features 8 DMRs (track ‘all_DMRs’). f-h, Enrichment analysis: depicted are –log10(qvalues)* sign(log_odds_ratio) from dark blue (depleted feature) to red (enriched feature) indicating the strength of enrichment of a given feature in a particular cluster as compared to all other clusters. (f) defined genomic regions, i.e. GENCODE transcription start sites (‘gencode_all_tss’), CpG islands (‘island’), CpG island shores (‘shore’), CpG island shelves (‘shelve’), intragenic regions (‘intragenic’), and intergenic regions (‘intergenic’), (g) hematopoietic enhancers as defined by published ChIP-seq data on primary murine hematopoietic cell types, and (h) transcription factor (TF) ChIP-seq peaks from hematopoietic cell types using the LOLA database (http://databio.org/regiondb).
