Hematopathologist-annotated mass cytometry dataset of acute myeloid leukemia diagnostic specimens

This dataset comprises hematopathologist-annotated single-cell data from patients with Acute Myeloid Leukemia (AML), sourced from FlowRepository (ID: FR-FCMZ2E7). The dataset includes normalized, singlet-gated events with annotations provided by the authorship team of Tsai et al., excluding one patient diagnosed with Myelodysplastic Syndrome (MDS) due to diagnostic ambiguity. To align with blast enumeration standards and pathology laboratory procedures, only CD45+ hematopoietic lineage cells were included in the analysis.
Data preprocessing followed standard mass cytometry protocols, including hyperbolic arcsine transformation with a cofactor of 5, scaling markers to their 99.9th percentile, and exclusion of cells with marker values exceeding this threshold to remove artifacts and outliers. This dataset, processed using the R package tidytof, offers a high-quality resource for researchers investigating AML through single-cell analysis., , , # Hematopathologist-annotated mass cytometry dataset of acute myeloid leukemia diagnostic specimens

[https://doi.org/10.5061/dryad.jq2bvq8kj](https://doi.org/10.5061/dryad.jq2bvq8kj)

## Description of the data and file structure

This dataset, originally collected to study single-cell morphometric profiling in AML, has now been annotated by a hematopathologist to provide single-cell labels distinguishing cancerous blasts from healthy bone marrow cells. These expert annotations, combined with the high-dimensional molecular and morphological data, make the dataset uniquely suited for developing and validating machine learning methods for disease-associated cell identification. The addition of cell-level annotations enables applications such as training classifiers to identify leukemia cells and selecting biologically meaningful features. This resource supports advances in single-cell data analysis and provides a valuable benchmark for studying acute myeloid leukemia and related conditio..., All human subjects data used in this study were obtained under protocols approved by the Institutional Review Board at Stanford University. Written informed consent was obtained from all participants, including consent for future research use and public data sharing of de-identified samples.

The data submitted to Dryad have been fully de-identified in accordance with HIPAA and international data protection guidelines. Specifically, all direct personal identifiers (e.g., names, dates of birth, medical record numbers) were removed prior to analysis. Clinical metadata were limited to non-identifiable variables.

As such, the dataset is compliant with applicable regulations governing the sharing of human subjects data and is suitable for public domain distribution via Dryad.