DNA Methylation Identifies Genetically and Prognostically Distinct Subtypes of Myelodysplastic Syndromes

We report the identification of five subtypes of myelodysplastic syndromes (MDS) identified via unsupervised classification of DNA methylation (DNAm) profiles. Bisulfite padlock probe sequencing (BSPP) was used to measure DNAm of ~500,000 unique CpGs covering 140,749 non-overlapping regulatory regions across the genome in bone marrow DNA samples from 141 patients with MDS. Application of a non-negative matrix factorization (NMF) decomposition of DNAm profiles identified five consensus clusters described by five NMF components as the most stable grouping solution. DNA methylomes were generated from bone marrow mononuclear cells of 141 myelodysplastic syndrome patients using targeted bisulfite sequencing approach, bisulfite padlock probes, and sequenced on an Illumina HiSeq 2500. Submitter declares that the raw data will be submitted to dbGaP (or EGA) because of patient privacy concerns.