"Plans DeepTrio trio and duo variant-calling workflows with child
Scanned 9/8/2026
Install to Claude Code
npx -y skills add VectorSpaceLab/AREX-Skill --skill trio-calling --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Trio Calling?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/vectorspacelab-trio-calling)More formats (shields.io, HTML) on the badges page.
---
name: trio-calling
description: "Plans DeepTrio trio and duo variant-calling workflows with child
and parent inputs, per-sample VCF/gVCF outputs, and GLnexus merge handoff."
disable-model-invocation: true
metadata:
disco-role: operating
license: BSD 3-Clause
---
# DeepTrio Trio Calling
Use this sub-skill when a user needs DeepTrio family calling for a child with one or two parents. It covers Docker/Bazel-style `run_deeptrio` command planning, duo behavior when one parent is absent, model-type choice, per-sample output planning, gVCF merge handoff, and family-specific troubleshooting.
## Route The Request
- Use this sub-skill for DeepTrio `run_deeptrio` workflows, trio/duo sample planning, child/parent output naming, gVCF planning, GLnexus handoff, and sex-chromosome parent-selection caveats.
- Route single-sample `run_deepvariant` germline calling to the `germline-calling` sub-skill.
- Route pangenome-aware DeepVariant workflows to the `pangenome-aware-calling` sub-skill.
- Route low-level `make_examples`, `call_variants`, `postprocess_variants`, sharded TFRecords, and extra-args internals to the `pipeline-stages` sub-skill.
- Route benchmarking reports, Mendelian-violation analysis, `hap.py`, RTG, and visualization to the `analysis-visualization` sub-skill.
## Fast Path
1. Identify family mode: trio has `child`, `parent1`, and `parent2`; duo has `child` and exactly one parent.
2. Choose `--model_type` from `WGS`, `WES`, `PACBIO`, or `ONT` according to the sequencing data and image/model availability.
3. Require an indexed FASTA and sorted, indexed, compatible BAMs for every supplied sample.
4. Require one VCF output for every supplied sample, and request a matching gVCF for every supplied sample when GLnexus or cohort merging is planned.
5. Use [`scripts/deeptrio_command_builder.py`](scripts/deeptrio_command_builder.py) to validate the plan and print a reviewable command without executing Docker, DeepTrio, or GLnexus.
6. Read [`references/workflows.md`](references/workflows.md), [`references/deeptrio-flags.md`](references/deeptrio-flags.md), and [`references/troubleshooting.md`](references/troubleshooting.md) for command patterns, flag contracts, and recovery steps.
## Bundled Helper
The helper is a command planner only:
```bash
python scripts/deeptrio_command_builder.py --help
```
It validates child/parent flag groups, makes duo mode explicit by omitting every parent2 flag, warns about WES regions and long-read candidate partitioning, and can emit a GLnexus handoff command when a complete gVCF set is present.
## Safety
Do not run Docker, GLnexus, network downloads, GPU jobs, or native DeepTrio tests without explicit user approval and real user-provided genomics inputs. Treat generated commands as plans until the user confirms the runtime environment, container mounts, and data paths.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!