tooluniverse-single-cell

When to Use This Skill

Apply when users:

• Have scRNA-seq data (h5ad, 10X, CSV count matrices) and want analysis
• Ask about cell type identification, clustering, or annotation
• Need differential expression analysis by cell type or condition
• Want gene-expression correlation analysis (e.g., gene length vs expression by cell type)
• Ask about PCA, UMAP, t-SNE for expression data
• Need Leiden/Louvain clustering on expression matrices
• Want statistical comparisons between cell types (t-test, ANOVA, fold change)
• Ask about marker genes for cell populations
• Need batch correction (Harmony, combat)
• Want trajectory or pseudotime analysis
• Ask about cell-cell communication (ligand-receptor interactions)
• Questions mention "single-cell", "scRNA-seq", "cell type", "h5ad"
• Questions involve immune cell types (CD4, CD8, CD14, CD19, monocytes, etc.)

BixBench Coverage: 18+ questions across 5 projects (bix-22, bix-27, bix-31, bix-33, bix-36)

NOT for (use other skills instead):

• Bulk RNA-seq DESeq2 analysis only → Use tooluniverse-rnaseq-deseq2
• Gene enrichment only (no expression data) → Use tooluniverse-gene-enrichment
• VCF/variant analysis → Use tooluniverse-variant-analysis
• Statistical modeling (regression, survival) → Use tooluniverse-statistical-modeling

---

Core Principles

1. Data-first approach - Load, inspect, and validate data before any analysis
2. AnnData-centric - All data flows through anndata objects for consistency
3. Cell type awareness - Many questions require per-cell-type subsetting and analysis
4. Statistical rigor - Proper normalization, multiple testing correction, effect sizes
5. Scanpy standard pipeline - Follow established best practices for scRNA-seq
6. Flexible input - Handle h5ad, 10X, CSV/TSV, pre-processed and raw data
7. Question-driven - Parse what the user is actually asking and extract the specific answer
8. Enrichment integration - Chain DE results into GO/KEGG/Reactome enrichment when requested
9. Large dataset support - Efficient handling of datasets with >100k cells

---

Optional

import harmonypy # batch correction bash pip install scanpy anndata leidenalg umap-learn harmonypy gseapy pandas numpy scipy scikit-learn statsmodels

---