nota-america/forgecat-agent-profiles / profiles/anthropics/knowledge-work-plugins/anthropics_knowledge-work-plugins_bio-research/for-forgecat/skills/scvi-tools
scvi-tools
VerifiedDeep learning for single-cell analysis using scvi-tools. This skill should be used when users need (1) data integration and batch correction with scVI/scANVI, (2) ATAC-seq analysis with PeakVI, (3) CITE-seq multi-modal analysis with totalVI, (4) multiome RNA+ATAC analysis with MultiVI, (5) spatial transcriptomics deconvolution with DestVI, (6) label transfer and reference mapping with scANVI/scArches, (7) RNA velocity with veloVI, or (8) any deep learning-based single-cell method. Triggers include mentions of scVI, scANVI, totalVI, PeakVI, MultiVI, DestVI, veloVI, sysVI, scArches, variational autoencoder, VAE, batch correction, data integration, multi-modal, CITE-seq, multiome, reference mapping, latent space.
| File | Kind | Size |
|---|---|---|
| LICENSE.txt | markdown | 10.0 KB |
| SKILL.md | skill md | 7.2 KB |
| references/atac_peakvi.md | markdown | 9.3 KB |
| references/batch_correction_sysvi.md | markdown | 10.2 KB |
| references/citeseq_totalvi.md | markdown | 10.1 KB |
| references/data_preparation.md | markdown | 6.2 KB |
| references/environment_setup.md | markdown | 6.3 KB |
| references/label_transfer.md | markdown | 10.2 KB |
| references/multiome_multivi.md | markdown | 8.9 KB |
| references/rna_velocity_velovi.md | markdown | 8.6 KB |
| references/scarches_mapping.md | markdown | 10.8 KB |
| references/scrna_integration.md | markdown | 9.6 KB |
| references/spatial_deconvolution.md | markdown | 11.2 KB |
| references/troubleshooting.md | markdown | 10.8 KB |
| scripts/cluster_embed.py | script | 6.3 KB |
| scripts/differential_expression.py | script | 6.7 KB |
| scripts/integrate_datasets.py | script | 6.9 KB |
| scripts/model_utils.py | script | 17.0 KB |
| scripts/prepare_data.py | script | 5.1 KB |
| scripts/train_model.py | script | 11.3 KB |
| scripts/transfer_labels.py | script | 7.2 KB |
| scripts/validate_adata.py | script | 11.9 KB |
scripts/prepare_data.py
View on GitHub#!/usr/bin/env python3
"""
Prepare AnnData for scvi-tools models.
This script handles QC filtering, HVG selection, and layer setup.
Output is ready for any scvi-tools model.
Usage:
python prepare_data.py input.h5ad output.h5ad --batch-key batch --n-hvgs 2000
python prepare_data.py input.h5ad output.h5ad --no-filter # Skip QC filtering
"""
import argparse
import sys
def prepare_data(
adata,
batch_key=None,
n_top_genes=2000,
min_genes=200,
max_genes=5000,
max_mito_pct=20.0,
min_cells=3,
skip_filter=False
):
"""
Prepare AnnData for scvi-tools.
Parameters
----------
adata : AnnData
Raw count data
batch_key : str, optional
Batch column for batch-aware HVG selection
n_top_genes : int
Number of highly variable genes
min_genes : int
Minimum genes per cell
max_genes : int
Maximum genes per cell
max_mito_pct : float
Maximum mitochondrial percentage
min_cells : int
Minimum cells per gene
skip_filter : bool
Skip QC filtering (use if already filtered)
Returns
-------
AnnData prepared for scvi-tools
"""
import scanpy as sc
import numpy as np
from model_utils import get_mito_genes
adata = adata.copy()
print(f"Input: {adata.shape[0]} cells, {adata.shape[1]} genes")
if not skip_filter:
# Calculate QC metrics
adata.var['mt'] = get_mito_genes(adata)
sc.pp.calculate_qc_metrics(adata, qc_vars=['mt'], inplace=True)
# Filter cells
n_before = adata.n_obs
adata = adata[adata.obs['n_genes_by_counts'] >= min_genes].copy()
adata = adata[adata.obs['n_genes_by_counts'] <= max_genes].copy()
adata = adata[adata.obs['pct_counts_mt'] < max_mito_pct].copy()
print(f"Filtered cells: {n_before} → {adata.n_obs}")
# Filter genes
n_genes_before = adata.n_vars
sc.pp.filter_genes(adata, min_cells=min_cells)
print(f"Filtered genes: {n_genes_before} → {adata.n_vars}")
# Store raw counts in layer
adata.layers["counts"] = adata.X.copy()
# HVG selection
if batch_key is not None and batch_key in adata.obs.columns:
print(f"Selecting {n_top_genes} HVGs (batch-aware: {batch_key})")
sc.pp.highly_variable_genes(
adata,
n_top_genes=n_top_genes,
flavor="seurat_v3",
batch_key=batch_key,
layer="counts"
)
else:
print(f"Selecting {n_top_genes} HVGs")
# Need to normalize for non-seurat_v3
sc.pp.normalize_total(adata, target_sum=1e4)
sc.pp.log1p(adata)
sc.pp.highly_variable_genes(adata, n_top_genes=n_top_genes)
# Restore counts to X
adata.X = adata.layers["counts"].copy()
# Subset to HVGs
n_hvg = adata.var['highly_variable'].sum()
adata = adata[:, adata.var['highly_variable']].copy()
print(f"Selected {n_hvg} highly variable genes")
print(f"Output: {adata.shape[0]} cells, {adata.shape[1]} genes")
return adata
def main():
parser = argparse.ArgumentParser(
description="Prepare AnnData for scvi-tools",
formatter_class=argparse.RawDescriptionHelpFormatter,
epilog="""
Examples:
# Basic preparation
python prepare_data.py raw.h5ad prepared.h5ad
# With batch-aware HVG selection
python prepare_data.py raw.h5ad prepared.h5ad --batch-key sample
# Custom parameters
python prepare_data.py raw.h5ad prepared.h5ad --n-hvgs 3000 --max-mito 15
# Skip filtering (data already QC'd)
python prepare_data.py filtered.h5ad prepared.h5ad --no-filter
"""
)
parser.add_argument("input", help="Input h5ad file")
parser.add_argument("output", help="Output h5ad file")
parser.add_argument("--batch-key", help="Batch column for HVG selection")
parser.add_argument("--n-hvgs", type=int, default=2000, help="Number of HVGs (default: 2000)")
parser.add_argument("--min-genes", type=int, default=200, help="Min genes per cell (default: 200)")
parser.add_argument("--max-genes", type=int, default=5000, help="Max genes per cell (default: 5000)")
parser.add_argument("--max-mito", type=float, default=20.0, help="Max mito %% (default: 20)")
parser.add_argument("--min-cells", type=int, default=3, help="Min cells per gene (default: 3)")
parser.add_argument("--no-filter", action="store_true", help="Skip QC filtering")
args = parser.parse_args()
try:
import scanpy as sc
except ImportError:
print("Error: scanpy required. Install with: pip install scanpy")
sys.exit(1)
# Load data
print(f"Loading {args.input}...")
adata = sc.read_h5ad(args.input)
# Prepare
adata = prepare_data(
adata,
batch_key=args.batch_key,
n_top_genes=args.n_hvgs,
min_genes=args.min_genes,
max_genes=args.max_genes,
max_mito_pct=args.max_mito,
min_cells=args.min_cells,
skip_filter=args.no_filter
)
# Save
print(f"Saving to {args.output}...")
adata.write_h5ad(args.output)
print("Done!")
if __name__ == "__main__":
main()
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |