Shaow/mouseskin_visium_wound

    # Mouse Skin Wound Healing Visium · POD0 + POD7

    Curated, ready-to-load spatial transcriptomics dataset.

    ## Source

    - Paper: [Foster et al., PNAS 2021](https://www.pnas.org/doi/10.1073/pnas.2110025118)
    - Canonical download: GEO GSE178758 · samples GSM6443210 (POD0) + GSM6443212 (POD7)

    ## Scale

    | Property | Value |
    |---|---|
    | Technology | 10x Genomics Visium |
    | Species | Mus musculus |
    | Tissue | Mouse skin (uninjured + post-operative day 7) |
    | Sections / slices | 2 |
    | Total cells / spots | 2,620 |

    ## Files

    - `h5ad/adata_skin_repair_pod0.h5ad`

• h5ad/adata_skin_repair_pod7.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)
  
    `pod`, `in_tissue`, `array_row`, `array_col`
  
    ## Notes
  
    Two Visium sections from a stented mouse-wound model: POD0 (uninjured baseline, 949 in-tissue spots × 32,285 genes) and POD7 (mid-repair, 1,671 spots × 32,285 genes). Spots already filtered to in_tissue==1.
  
    ## Usage
  
    ```python
  

import scanpy as sc from huggingface_hub import snapshot_download d = snapshot_download(repo_id='Shaow/mouseskin_visium_wound', repo_type='dataset') adata_pod0 = sc.read_h5ad(f'{d}/h5ad/adata_skin_repair_pod0.h5ad') adata_pod7 = sc.read_h5ad(f'{d}/h5ad/adata_skin_repair_pod7.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.