Curated Age-Dependent AML Transcriptome Datasets

The curation method is summarized in the Supplementary File S3 flowchart and in the Results section. For our analysis we included 34 age-dependent datasets from 32 different studies, 16 included AML and 18 healthy subjects respectively. From the 34 datasets, 32 were produced from Affymetrix GeneChip Human Genome U133 Plus 2.0 (GPL570) and 2 conducted on Affymetrix GeneChip Human Genome U133 Array Set (GPL96 & GPL97) arrays. Table 1 provides detailed information about each dataset, including the number of samples used from each dataset, sample tissue source, as well as the total number of AML patients and healthy subjects. Two studies, GSE12417^{55 (link)} and GSE37642^{56 (link)–59 (link)}, were originally conducted on two different Affymetrix array types (GPL570, and GPL96 & GPL97), so each was separated into two subgroups and each subgroup was considered as individual dataset in our analysis, dataset GSE12417: (i) subgroup 1 included 73 BM and 5 PB samples, and (ii) subgroup 2 included 160 BM and 2 PB. For dataset GSE37642: (i) subgroup 1 included 140 BM and (ii) subgroup 2 included 422 BM samples (Table 1).