alterlab-scvi-tools
Train deep generative models for single-cell omics with scvi-tools — probabilistic batch correction and integration (scVI), reference-mapping transfer learning (scArches), differential expression with uncertainty, and multimodal models (totalVI for CITE-seq, MultiVI for multiome)
Install
npx skills add https://github.com/AlterLab-IEU/AlterLab-Academic-Skills/tree/main/skills/bioinformatics/alterlab-scvi-tools
claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install alterlab-ieu-alterlab-academic-skills@llmmart
git clone https://github.com/AlterLab-IEU/AlterLab-Academic-Skills.git
The skills CLI installs just this skill, for any of its supported agents. Claude Code installs the whole alterlab-ieu/alterlab-academic-skills collection as a plugin from our marketplace. Git is the plain clone.
Skill manifest
scvi-tools
Overview
scvi-tools is a comprehensive Python framework for probabilistic models in single-cell genomics. Built on PyTorch and PyTorch Lightning, it provides deep generative models using variational inference for analyzing diverse single-cell data modalities.
When to Use This Skill
Use this skill when:
- Analyzing single-cell RNA-seq data (dimensionality reduction, batch correction, integration)
- Working with single-cell ATAC-seq or chromatin accessibility data
- Integrating multimodal data (CITE-seq, multiome, paired/unpaired datasets)
- Analyzing spatial transcriptomics data (deconvolution, spatial mapping)
- Performing differential expression analysis on single-cell data
- Conducting cell type annotation or transfer learning tasks
- Working with specialized single-cell modalities (methylation, cytometry, RNA velocity)
- Building custom probabilistic models for single-cell analysis
Does NOT Trigger
| Scenario | Use Instead |
|---|---|
| Standard QC -> normalize -> HVG -> PCA -> UMAP -> Leiden -> marker DE | alterlab-scanpy |
Reading, reshaping, or concatenating the .h5ad object itself |
alterlab-anndata |
| Zero-shot annotation or embeddings from a pretrained foundation model | alterlab-scgpt |
| RNA velocity from spliced/unspliced counts (scVelo's EM/dynamical models) | alterlab-scvelo |
| Spatial neighborhood statistics, Moran's I, ligand-receptor on a Visium/Xenium slide | alterlab-squidpy-spatial |
| Bulk RNA-seq differential expression | alterlab-pydeseq2 |
Hand-off rule. When the request matches a row above, open your answer by naming that skill by its exact id (for example alterlab-scanpy) and saying in one line why it fits better, then stop applying this skill's procedure. Name it even when you also need to ask for missing input. If only part of the request is out of scope, name the skill for that part and do only the in-scope part. A brief general pointer is fine; the named skill does the work.
Core Capabilities
scvi-tools provides models organized by data modality:
1. Single-Cell RNA-seq Analysis
Core models for expression analysis, batch correction, and integration. See references/models-scrna-seq.md for:
- scVI: Unsupervised dimensionality reduction and batch correction
- scANVI: Semi-supervised cell type annotation and integration
- AUTOZI: Zero-inflation detection and modeling
- VeloVI: RNA velocity analysis
- contrastiveVI: Perturbation effect isolation
2. Chromatin Accessibility (ATAC-seq)
Models for analyzing single-cell chromatin data. See references/models-atac-seq.md for:
- PeakVI: Peak-based ATAC-seq analysis and integration
- PoissonVI: Quantitative fragment count modeling
- scBasset: Deep learning approach with motif analysis
3. Multimodal & Multi-omics Integration
Joint analysis of multiple data types. See references/models-multimodal.md for:
- totalVI: CITE-seq protein and RNA joint modeling
- MultiVI: Paired and unpaired multi-omic integration
- MrVI: Multi-resolution cross-sample analysis
4. Spatial Transcriptomics
Spatially-resolved transcriptomics analysis. See references/models-spatial.md for:
- DestVI: Multi-resolution spatial deconvolution
- Stereoscope: Cell type deconvolution
- Tangram: Spatial mapping and integration
- scVIVA: Cell-environment relationship analysis
5. Specialized Modalities
Additional specialized analysis tools. See references/models-specialized.md for:
- MethylVI/MethylANVI: Single-cell methylation analysis
- CytoVI: Flow/mass cytometry batch correction
- Solo: Doublet detection
- CellAssign: Marker-based cell type annotation
Typical Workflow
All scvi-tools models follow a consistent API pattern:
# 1. Load and preprocess data (AnnData format)
import scvi
import scanpy as sc
adata = scvi.data.heart_cell_atlas_subsampled()
sc.pp.filter_genes(adata, min_counts=3)
sc.pp.highly_variable_genes(adata, n_top_genes=1200)
# 2. Register data with model (specify layers, covariates)
scvi.model.SCVI.setup_anndata(
adata,
layer="counts", # Use raw counts, not log-normalized
batch_key="batch",
categorical_covariate_keys=["donor"],
continuous_covariate_keys=["percent_mito"]
)
# 3. Create and train model
model = scvi.model.SCVI(adata)
model.train()
# 4. Extract latent representations and normalized values
latent = model.get_latent_representation()
normalized = model.get_normalized_expression(library_size=1e4)
# 5. Store in AnnData for downstream analysis
adata.obsm["X_scVI"] = latent
adata.layers["scvi_normalized"] = normalized
# 6. Downstream analysis with scanpy
sc.pp.neighbors(adata, use_rep="X_scVI")
sc.tl.umap(adata)
sc.tl.leiden(adata)
Key Design Principles:
- Raw counts required: Models expect unnormalized count data for optimal performance
- Unified API: Consistent interface across all models (setup → train → extract)
- AnnData-centric: Seamless integration with the scanpy ecosystem
- GPU acceleration: Automatic utilization of available GPUs
- Batch correction: Handle technical variation through covariate registration
Common Analysis Tasks
Differential Expression
Probabilistic DE analysis using the learned generative models:
de_results = model.differential_expression(
groupby="cell_type",
group1="TypeA",
group2="TypeB",
mode="change", # composite hypothesis testing with an effect-size threshold
delta=0.25, # minimum |LFC| to count as a real change
)
# Significance lives in proba_de (posterior prob. of DE), NOT a p-value.
# In "change" mode this is the posterior prob. that |LFC| > delta.
sig = de_results[de_results["proba_de"] > 0.95]
See references/differential-expression.md for the full output schema and interpretation.
When the DE runs from a script, quote the model.differential_expression(...) call in the reply with every argument you set (groupby, group1, group2, mode, delta, batch_correction, fdr_target), plus the proba_de cutoff and the paths of the files written. Those settings define the test; a methods section and anyone re-running it need them verbatim, not paraphrased.
Model Persistence
Save and load trained models:
# Save model
model.save("./model_directory", overwrite=True)
# Load model
model = scvi.model.SCVI.load("./model_directory", adata=adata)
Batch Correction and Integration
Integrate datasets across batches or studies:
# Register batch information
scvi.model.SCVI.setup_anndata(adata, batch_key="study")
# Model automatically learns batch-corrected representations
model = scvi.model.SCVI(adata)
model.train()
latent = model.get_latent_representation() # Batch-corrected
Theoretical Foundations
scvi-tools is built on:
- Variational inference: Approximate posterior distributions for scalable Bayesian inference
- Deep generative models: VAE architectures that learn complex data distributions
- Amortized inference: Shared neural networks for efficient learning across cells
- Probabilistic modeling: Principled uncertainty quantification and statistical testing
See references/theoretical-foundations.md for detailed background on the mathematical framework.
Additional Resources
- Workflows:
references/workflows.mdcontains common workflows, best practices, hyperparameter tuning, and GPU optimization - Model References: Detailed documentation for each model category in the
references/directory - Official Documentation: https://docs.scvi-tools.org/en/stable/
- Tutorials: https://docs.scvi-tools.org/en/stable/tutorials/index.html
- API Reference: https://docs.scvi-tools.org/en/stable/api/index.html
Installation
scvi-tools 1.5 requires Python >= 3.12 (PyPI requires_python, checked 2026-11-04 for
1.5.2), so pin the interpreter. Run install-free with uv, which resolves the package per run (a
bare uv pip install fails outside a virtual environment, and a plain uv run python may pick an
older Python):
uv run --python ">=3.12" --with scvi-tools --with scanpy python analysis.py
uv run --python ">=3.12" --with "scvi-tools[cuda]" python analysis.py # CUDA 12; [cuda13] for CUDA 13
uv run --python ">=3.12" --with "scvi-tools[metal]" python analysis.py # Apple silicon (MPS)
uv run --python ">=3.12" --with "scvi-tools[autotune]" python analysis.py # Ray Tune search
In a project with its own virtual environment, uv pip install scvi-tools into that
environment instead.
Version note (2026-09): scvi-tools 1.5 dropped the JAX backend entirely, so scvi.model.JaxSCVI
and the other Jax-prefixed models no longer exist — everything runs on PyTorch/Lightning (with an
mlxSCVI variant for Apple silicon). Tangram also moved to a PyTorch backend in 1.5.0, so
results can differ slightly from JAX-era runs; re-run rather than mixing old and new outputs in
one comparison.
Best Practices
- Use raw counts: Always provide unnormalized count data to models
- Filter genes: Remove low-count genes before analysis (e.g.,
min_counts=3) - Register covariates: Include known technical factors (batch, donor, etc.) in
setup_anndata - Feature selection: Use highly variable genes for improved performance
- Model saving: Always save trained models to avoid retraining
- GPU usage: Enable GPU acceleration for large datasets (
accelerator="gpu") - Scanpy integration: Store outputs in AnnData objects for downstream analysis
Files (alterlab-academic-skills)
-
evals
-
evals.json 4.3 KB
Loading…
-
-
references
-
differential-expression.md 14.8 KB
Loading… -
models-atac-seq.md 9.4 KB
Loading… -
models-multimodal.md 10.2 KB
Loading… -
models-scrna-seq.md 9.6 KB
Loading… -
models-spatial.md 11 KB
Loading… -
models-specialized.md 10.2 KB
Loading… -
theoretical-foundations.md 11.4 KB
Loading… -
workflows.md 13 KB
Loading…
-
-
SKILL.md 10.4 KB
--- name: alterlab-scvi-tools description: Train deep generative models for single-cell omics with scvi-tools — probabilistic batch correction and integration (scVI), reference-mapping transfer learning (scArches), differential expression with uncertainty, and multimodal models (totalVI for CITE-seq, MultiVI for multiome). Use when correcting batch effects, integrating multimodal data, or doing advanced probabilistic single-cell modeling — for standard analysis pipelines use scanpy. Part of the AlterLab Academic Skills suite. license: MIT allowed-tools: Read Write Edit Bash(python:*) Bash(uv:*) compatibility: "Self-contained — runs under `uv run --python \">=3.12\" --with scvi-tools`; no API key or account required. Written for scvi-tools 1.5 (current 1.5.1 as of 2026-09), which requires Python >= 3.12. A GPU is optional but makes training practical on large datasets." metadata: skill-author: AlterLab version: "1.1.2" last_updated: "2026-11-04" --- # scvi-tools ## Overview scvi-tools is a comprehensive Python framework for probabilistic models in single-cell genomics. Built on PyTorch and PyTorch Lightning, it provides deep generative models using variational inference for analyzing diverse single-cell data modalities. ## When to Use This Skill Use this skill when: - Analyzing single-cell RNA-seq data (dimensionality reduction, batch correction, integration) - Working with single-cell ATAC-seq or chromatin accessibility data - Integrating multimodal data (CITE-seq, multiome, paired/unpaired datasets) - Analyzing spatial transcriptomics data (deconvolution, spatial mapping) - Performing differential expression analysis on single-cell data - Conducting cell type annotation or transfer learning tasks - Working with specialized single-cell modalities (methylation, cytometry, RNA velocity) - Building custom probabilistic models for single-cell analysis ### Does NOT Trigger | Scenario | Use Instead | |----------|-------------| | Standard QC -> normalize -> HVG -> PCA -> UMAP -> Leiden -> marker DE | `alterlab-scanpy` | | Reading, reshaping, or concatenating the `.h5ad` object itself | `alterlab-anndata` | | Zero-shot annotation or embeddings from a pretrained foundation model | `alterlab-scgpt` | | RNA velocity from spliced/unspliced counts (scVelo's EM/dynamical models) | `alterlab-scvelo` | | Spatial neighborhood statistics, Moran's I, ligand-receptor on a Visium/Xenium slide | `alterlab-squidpy-spatial` | | Bulk RNA-seq differential expression | `alterlab-pydeseq2` | **Hand-off rule.** When the request matches a row above, open your answer by naming that skill by its exact id (for example `alterlab-scanpy`) and saying in one line why it fits better, then stop applying this skill's procedure. Name it even when you also need to ask for missing input. If only part of the request is out of scope, name the skill for that part and do only the in-scope part. A brief general pointer is fine; the named skill does the work. ## Core Capabilities scvi-tools provides models organized by data modality: ### 1. Single-Cell RNA-seq Analysis Core models for expression analysis, batch correction, and integration. See `references/models-scrna-seq.md` for: - **scVI**: Unsupervised dimensionality reduction and batch correction - **scANVI**: Semi-supervised cell type annotation and integration - **AUTOZI**: Zero-inflation detection and modeling - **VeloVI**: RNA velocity analysis - **contrastiveVI**: Perturbation effect isolation ### 2. Chromatin Accessibility (ATAC-seq) Models for analyzing single-cell chromatin data. See `references/models-atac-seq.md` for: - **PeakVI**: Peak-based ATAC-seq analysis and integration - **PoissonVI**: Quantitative fragment count modeling - **scBasset**: Deep learning approach with motif analysis ### 3. Multimodal & Multi-omics Integration Joint analysis of multiple data types. See `references/models-multimodal.md` for: - **totalVI**: CITE-seq protein and RNA joint modeling - **MultiVI**: Paired and unpaired multi-omic integration - **MrVI**: Multi-resolution cross-sample analysis ### 4. Spatial Transcriptomics Spatially-resolved transcriptomics analysis. See `references/models-spatial.md` for: - **DestVI**: Multi-resolution spatial deconvolution - **Stereoscope**: Cell type deconvolution - **Tangram**: Spatial mapping and integration - **scVIVA**: Cell-environment relationship analysis ### 5. Specialized Modalities Additional specialized analysis tools. See `references/models-specialized.md` for: - **MethylVI/MethylANVI**: Single-cell methylation analysis - **CytoVI**: Flow/mass cytometry batch correction - **Solo**: Doublet detection - **CellAssign**: Marker-based cell type annotation ## Typical Workflow All scvi-tools models follow a consistent API pattern: ```python # 1. Load and preprocess data (AnnData format) import scvi import scanpy as sc adata = scvi.data.heart_cell_atlas_subsampled() sc.pp.filter_genes(adata, min_counts=3) sc.pp.highly_variable_genes(adata, n_top_genes=1200) # 2. Register data with model (specify layers, covariates) scvi.model.SCVI.setup_anndata( adata, layer="counts", # Use raw counts, not log-normalized batch_key="batch", categorical_covariate_keys=["donor"], continuous_covariate_keys=["percent_mito"] ) # 3. Create and train model model = scvi.model.SCVI(adata) model.train() # 4. Extract latent representations and normalized values latent = model.get_latent_representation() normalized = model.get_normalized_expression(library_size=1e4) # 5. Store in AnnData for downstream analysis adata.obsm["X_scVI"] = latent adata.layers["scvi_normalized"] = normalized # 6. Downstream analysis with scanpy sc.pp.neighbors(adata, use_rep="X_scVI") sc.tl.umap(adata) sc.tl.leiden(adata) ``` **Key Design Principles:** - **Raw counts required**: Models expect unnormalized count data for optimal performance - **Unified API**: Consistent interface across all models (setup → train → extract) - **AnnData-centric**: Seamless integration with the scanpy ecosystem - **GPU acceleration**: Automatic utilization of available GPUs - **Batch correction**: Handle technical variation through covariate registration ## Common Analysis Tasks ### Differential Expression Probabilistic DE analysis using the learned generative models: ```python de_results = model.differential_expression( groupby="cell_type", group1="TypeA", group2="TypeB", mode="change", # composite hypothesis testing with an effect-size threshold delta=0.25, # minimum |LFC| to count as a real change ) # Significance lives in proba_de (posterior prob. of DE), NOT a p-value. # In "change" mode this is the posterior prob. that |LFC| > delta. sig = de_results[de_results["proba_de"] > 0.95] ``` See `references/differential-expression.md` for the full output schema and interpretation. When the DE runs from a script, quote the `model.differential_expression(...)` call in the reply with every argument you set (`groupby`, `group1`, `group2`, `mode`, `delta`, `batch_correction`, `fdr_target`), plus the `proba_de` cutoff and the paths of the files written. Those settings define the test; a methods section and anyone re-running it need them verbatim, not paraphrased. ### Model Persistence Save and load trained models: ```python # Save model model.save("./model_directory", overwrite=True) # Load model model = scvi.model.SCVI.load("./model_directory", adata=adata) ``` ### Batch Correction and Integration Integrate datasets across batches or studies: ```python # Register batch information scvi.model.SCVI.setup_anndata(adata, batch_key="study") # Model automatically learns batch-corrected representations model = scvi.model.SCVI(adata) model.train() latent = model.get_latent_representation() # Batch-corrected ``` ## Theoretical Foundations scvi-tools is built on: - **Variational inference**: Approximate posterior distributions for scalable Bayesian inference - **Deep generative models**: VAE architectures that learn complex data distributions - **Amortized inference**: Shared neural networks for efficient learning across cells - **Probabilistic modeling**: Principled uncertainty quantification and statistical testing See `references/theoretical-foundations.md` for detailed background on the mathematical framework. ## Additional Resources - **Workflows**: `references/workflows.md` contains common workflows, best practices, hyperparameter tuning, and GPU optimization - **Model References**: Detailed documentation for each model category in the `references/` directory - **Official Documentation**: https://docs.scvi-tools.org/en/stable/ - **Tutorials**: https://docs.scvi-tools.org/en/stable/tutorials/index.html - **API Reference**: https://docs.scvi-tools.org/en/stable/api/index.html ## Installation scvi-tools 1.5 requires **Python >= 3.12** (PyPI `requires_python`, checked 2026-11-04 for 1.5.2), so pin the interpreter. Run install-free with uv, which resolves the package per run (a bare `uv pip install` fails outside a virtual environment, and a plain `uv run python` may pick an older Python): ```bash uv run --python ">=3.12" --with scvi-tools --with scanpy python analysis.py uv run --python ">=3.12" --with "scvi-tools[cuda]" python analysis.py # CUDA 12; [cuda13] for CUDA 13 uv run --python ">=3.12" --with "scvi-tools[metal]" python analysis.py # Apple silicon (MPS) uv run --python ">=3.12" --with "scvi-tools[autotune]" python analysis.py # Ray Tune search ``` In a project with its own virtual environment, `uv pip install scvi-tools` into that environment instead. Version note (2026-09): scvi-tools 1.5 dropped the JAX backend entirely, so `scvi.model.JaxSCVI` and the other Jax-prefixed models no longer exist — everything runs on PyTorch/Lightning (with an `mlxSCVI` variant for Apple silicon). `Tangram` also moved to a PyTorch backend in 1.5.0, so results can differ slightly from JAX-era runs; re-run rather than mixing old and new outputs in one comparison. ## Best Practices 1. **Use raw counts**: Always provide unnormalized count data to models 2. **Filter genes**: Remove low-count genes before analysis (e.g., `min_counts=3`) 3. **Register covariates**: Include known technical factors (batch, donor, etc.) in `setup_anndata` 4. **Feature selection**: Use highly variable genes for improved performance 5. **Model saving**: Always save trained models to avoid retraining 6. **GPU usage**: Enable GPU acceleration for large datasets (`accelerator="gpu"`) 7. **Scanpy integration**: Store outputs in AnnData objects for downstream analysis
Comments (0)
Sign in to join the conversation.
Reviews (0)
No reviews yet.
No comments yet.