Fig. 1

Heterogenous 5hmC distributions in patients with myeloid neoplasms. A The experimental design for genome-wide sCMS-IP-seq analysis in the cohort. B Boxplot representation of the identified 5hmC peak numbers in the analyzed cohort. Bounds of the box span from 25 to 75% percentile, and the center line within each box represents the median. Whiskers represent median ± 1.5 times interquartile range. C Heatmap representation of the enrichment of selected 5hmC peaks among all the analyzed cohorts. The selected 5hmC peaks (n = 12,540) exhibited small variations (coefficient of variation, CV < 0.15) in healthy donor groups, whereas MDS and cancer patients showed higher variation. Each row represents a selected 5hmC peak, each column represents an individual person. RAEB: Refractory anemia with excess blasts; RARS: Refractory anemia with ring sideroblasts; RCMD: Refractory cytopenia with multilineage dysplasia; RS: Ring sideroblasts. D Top 10 selected categories of the GREAT analysis results for the 5hmC peaks shown in Fig. 1C. E Heatmap representation of the top 20,000 variable disease-specific differentially hydroxymethylated regions (DHMRs). F Box-plot showing the enrichment of 5hmC within published DNaseI hypersensitive sites, H3K4me1-, or H3K27ac-enriched genomic regions in healthy controls, as well as patients diagnosed with myeloid neoplasms. Bounds of the box span from 25 to 75% percentile, the center line within the box represents the median. Whiskers represent median ± 1.5 times interquartile range. G The t-SNE plot of the DNA methylation level within disease-specific DHMRs in the published AML cohort. Hypo-DHMRs: genomic regions showed decreased 5hmC levels in patients compared with healthy controls; Hyper-DHMRs: genomic regions showed increased 5hmC levels in patients compared with healthy controls. The DNA methylation levels at each CpG site were obtained from published datasets collected from CD34 + cells in healthy control (phs000159) or bone marrow aspirates in AML patients (GSE98350)