##### Scanpy analysis

Agentic re-implementation of the classic PBMC3k Scanpy tutorial: QC,
clustering, marker discovery, and cell-type annotation — with cutoffs
and labels chosen from the data, not hard-coded from the docs.

**Instance:** laminlabs/lamindata

#### Agentic analysis

| --- | --- | --- |
| Step | Transform | Outputs |
| =================================== | =================================== | =================================== |
| Preprocess, cluster, annotate | clustering.py · agent run | annotated h5ad · cluster markers |
| · QC violins · UMAP · marker |
| dotplot |
| --- | --- | --- |

The cell-type rationale (QC cutoffs, cluster→type map, caveats) is
attached as notes on the annotated AnnData.