Deep generative models for single-cell omics — probabilistic batch correction, transfer learning, differential expression with uncertainty, and multi-modal integration.
Wraps the scvi-tools library for advanced single-cell analysis: scVI for batch correction and latent space modeling, TOTALVI for CITE-seq (RNA + protein), MultiVI for multiome (RNA + ATAC), scANVI for semi-supervised cell type annotation, and differential expression with posterior uncertainty quantification.
cd scvi-tools
python3 -m venv .venv && source .venv/bin/activate && pip install scvi-tools -qNone.
python3 scripts/demo.py --format summaryscvi-tools
- Direct script run: pass (--help)
- Agno agent (Claude Haiku 4.5): pass
Agent loaded the skill and described scVI, TOTALVI, MultiVI, and scANVI models with their respective use cases.
- Cleaned
__pycache__/, 1 ruff lint fix