Babysitter deepvariant-caller
DeepVariant deep learning variant calling skill for high-accuracy SNV and indel detection
install
source · Clone the upstream repo
git clone https://github.com/a5c-ai/babysitter
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/a5c-ai/babysitter "$T" && mkdir -p ~/.claude/skills && cp -r "$T/library/specializations/domains/science/bioinformatics/skills/deepvariant-caller" ~/.claude/skills/a5c-ai-babysitter-deepvariant-caller && rm -rf "$T"
manifest:
library/specializations/domains/science/bioinformatics/skills/deepvariant-caller/SKILL.mdsource content
DeepVariant Caller Skill
Purpose
Enable DeepVariant deep learning variant calling for high-accuracy SNV and indel detection.
Capabilities
- GPU-accelerated variant calling
- WGS/WES/PacBio mode selection
- Model customization and retraining
- Confidence calibration
- Multi-sample variant calling
- Docker/Singularity deployment
Usage Guidelines
- Select appropriate model for sequencing type
- Use GPU acceleration when available
- Validate accuracy against benchmark datasets
- Consider container deployment for reproducibility
- Document model version and parameters
- Compare with traditional callers for validation
Dependencies
- DeepVariant
- Parabricks
Process Integration
- Whole Genome Sequencing Pipeline (wgs-analysis-pipeline)
- Long-Read Sequencing Analysis (long-read-analysis)
- Analysis Pipeline Validation (pipeline-validation)