---
license: mit
language:
- en
tags:
- spatial-transcriptomics
- snrna-seq
- mus
pretty_name: Mouse Brain snRNA-seq Reference (cell2location)
size_categories:
- 10K<n<100K
---

        # Mouse Brain snRNA-seq Reference (cell2location)

        Curated, ready-to-load spatial transcriptomics dataset.

        ## Source

        - Paper: [Kleshchevnikov et al., Nat. Biotechnol. 2022 (cell2location)](https://www.nature.com/articles/s41587-021-01139-4)
        - Canonical download: cell2location.cog.sanger.ac.uk/tutorial/mouse_brain_snrna/all_cells_20200625.h5ad

        ## Scale

        | Property | Value |
        |---|---|
        | Technology | snRNA-seq (10x Chromium) |
        | Species | Mus musculus |
        | Tissue | Mouse brain |
        | Sections / slices | 0 |
        | Total cells / spots | 40,572 |

        ## Files

        - `all_cells_20200625.h5ad`

        Each `.h5ad` follows the AnnData spec:
        - `.X` — gene expression matrix (cells × genes), sparse where natural
        - `.obs` — per-cell annotations (see "Metadata" below)
        - `.obsm['spatial']` — `(n_cells, 2)` float32 spatial coordinates
        - (where present) `.layers['count']` — raw integer counts
        - (where present) `.obsm['spatial3d']` — `(n_cells, 3)` float32 (x, y, z=section)

        ## Metadata (`obs` columns)

        `sample`, `barcode`

        ## Notes

        Mouse-brain snRNA-seq reference from the cell2location tutorial. 40,572 nuclei × 31,053 genes, raw counts. Useful as an external single-cell atlas for cell-type annotation of MERFISH or Slide-seqV2 spatial data.

        ## Usage

        ```python
import scanpy as sc
from huggingface_hub import snapshot_download
d = snapshot_download(repo_id='Shaow/mousebrain_snrnaseq_cell2location', repo_type='dataset')
adata_ref = sc.read_h5ad(f'{d}/all_cells_20200625.h5ad')
```


        ## Citation

        If you use this dataset, please cite the source paper above.

        ## License

        MIT for the curation/preparation. Underlying data inherits the license of
        the upstream publication (typically CC-BY-4.0); please see the source paper.