single-cell-rnaseq

name	single-cell-rnaseq
description	Single-cell RNA-seq analysis (mirrors Claude Science's single-cell skill). scanpy/anndata pipeline QC->normalize->HVG->PCA->neighbors->Leiden->UMAP->markers->annotation, publication figures. Use for scRNA-seq, count matrices, cell clustering/annotation. Needs the sci-bio env.

single-cell-rnaseq

Depends on scanpy, anndata (see the sci-bio env in environment-bio.yml). Gene annotation via connectors/ensembl.py / uniprot.py.

Standard workflow (write the script to runs//outputs/sc.py, fix the seed)

1. Read count matrix (inputs/) -> AnnData.
2. QC: pct_counts_mt, n_genes, doublets; record filter thresholds and pre/post cell counts.
3. normalize_total + log1p + highly_variable_genes + scale.
4. pca -> neighbors -> leiden -> umap.
5. rank_genes_groups for markers -> annotate cell types (state the basis).
6. Figures (UMAP, marker dotplot) via figure-maker, saved to outputs/.

Hard rules

• Fix and record the random seed; every annotation traces back to markers and parameters.