README.md

---
dataset_name: ainciburu2023_hematopoiesis_aging_mds
annotations_creators:
  - expert-generated
language:
  - en
multilinguality: "no"
pretty_name: Single-cell Profiling of Human Hematopoiesis Across Aging and Myeloid Malignancies (Ainciburu et al. 2023)
task_categories:
  - other
size_categories:
  - 100K<n<1M
license: cc-by-4.0
dataset_type: biomedical
---


# Dataset Card for `ainciburu_processed`

## Dataset Summary

This dataset comprises ~115,000 CD34+ hematopoietic stem and progenitor cells (HSPCs) profiled via 10x Genomics single-cell RNA-seq from healthy young, elderly, and myelodysplastic syndrome (MDS) patients. The data originates from:

> *Uncovering perturbations in human hematopoiesis associated with healthy aging and myeloid malignancies at single-cell resolution*  
> — Ainciburu et al., *eLife* (2023)  
> [PMCID: PMC9904760](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9904760/)  
> [DOI: 10.7554/eLife.79363](https://doi.org/10.7554/eLife.79363)

## Transformation Summary

The raw files were processed using the following pipeline:

1. **Data Acquisition**:
   - Extracted from GEO accession `GSE180298`, including raw matrix `.h5` files and associated metadata (`*_metadata.txt.gz`).

2. **Data Parsing and Merging**:
   - Read individual 10x `.h5` matrices per sample.
   - Assigned unique cell barcodes including the sample identifier.
   - Merged all into a single `AnnData` object with batch annotations.

3. **Metadata Alignment**:
   - Loaded metadata for young, elderly, and MDS samples.
   - Used barcode-sample combinations to merge metadata with cell observations.
   - Mapped `cell_type`, `patient_id`, and manually defined `patient_age`.

4. **Final Output**:
   - Saved unified dataset as `processed/ainciburu_processed.h5ad`.

## Supported Tasks and Benchmarks

- **Aging and Disease Comparison**: Track HSPC shifts from youth to old age and MDS.
- **Trajectory Inference**: Includes pseudotime, lineage tracing via STREAM and Palantir.
- **GRN Analysis**: SCENIC-based regulon detection per age/disease group.
- **Subtype Classification**: Includes supervised label transfer from healthy to diseased donors.
- **Differential Expression and GSEA**: Precomputed per lineage and age/disease state.

## Languages

All annotations and metadata are in English.

## Dataset Structure

### Data Instances

Each row is a single cell with:

- Raw gene expression (UMIs)
- Cell type annotation (e.g., HSC, MEP, GMP)
- Sample-level metadata (patient ID, condition, age)
- Batch label (sample ID)

### Data Splits

No formal splits; users may stratify by:

- `patient_id`
- `patient_age` (continuous)
- `condition`: `"young"`, `"elderly"`, `"mds"`

## Dataset Creation

### Curation Rationale

Aimed to reveal regulatory, transcriptional, and population-level changes in hematopoiesis across the lifespan and in disease (MDS), using high-resolution single-cell RNA-seq and computational modeling.

### Source Data

- Bone marrow CD34+ cells from 5 young (19–23y), 3 elderly (61–74y), and 4 MDS patients (54–83y).
- Sorted, sequenced using 10x Genomics Chromium.
- Raw data from GEO accession `GSE180298`.

### Preprocessing Details

- Metadata aligned by barcode + sample combination
- Cell type labels derived from supervised classification and manual annotation
- Cell-level age assignments based on patient identity
- Final object stored as a single `.h5ad` file for interoperability

## Licensing Information

This dataset is released under the **Creative Commons BY 4.0** license. Please cite the original publication when using this dataset in your work.

## Citation

```bibtex
@article{ainciburu2023aging,
  title={Uncovering perturbations in human hematopoiesis associated with healthy aging and myeloid malignancies at single-cell resolution},
  author={Ainciburu, Marina and Ezponda, Teresa and Berastegui, Nerea and others},
  journal={eLife},
  volume={12},
  pages={e79363},
  year={2023},
  publisher={eLife Sciences Publications Limited},
  doi={10.7554/eLife.79363}
}