This includes processed scRNA-seq profiles from the 21 AML studies included in the analysis of Zeng et al. Blood Cancer Discovery 2025 (DOI: 10.1158/2643-3230.BCD-24-0342). For each of these studies, single-cell transcriptomes were projected onto BoneMarrowMap, followed by cell-type classification, pseudotime projections, and scoring of various leukemia stem cell (LSC)-related signatures. This amounts to approximately ~1.2 million leukemia cells and an additional ~600,000 non-leukemic immune cells. Some healthy control samples have also been included within this dataset. Separate RDS files are provided for each study, but the metadata annotations and UMAP projections are uniformly generated. A CSV file has been provided with metadata for all ~1.8 million cells spanning the 21 AML scRNA-seq studies, and descriptions for each metadata column are included as a separate XLSX file. Please see the Blood Cancer Discovery paper for references to each original study from which each scRNA-seq dataset is derived. For more information on the BoneMarrowMap reference atlas of scRNA-seq across human hematopoietic differentiation, please see below: CellxGene Dataset: https://cellxgene.cziscience.com/collections/f6c50495-3361-40ed-a819-fb9644396ed9 Github R Package: https://github.com/andygxzeng/BoneMarrowMap