# Bio Binning Qc

> Perform metagenomic binning with QuickBin, refinement, and QC with completeness/contamination checks.

- **Type:** Skill
- **Install:** `agentstack add skill-fmschulz-omics-skills-bio-binning-qc`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [fmschulz](https://agentstack.voostack.com/s/fmschulz)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [fmschulz](https://github.com/fmschulz)
- **Source:** https://github.com/fmschulz/omics-skills/tree/main/skills/bio-binning-qc
- **Website:** https://fmschulz.github.io/omics-skills/

## Install

```sh
agentstack add skill-fmschulz-omics-skills-bio-binning-qc
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# Bio Binning QC

Perform metagenomic binning, refinement, and QC with completeness/contamination checks.

## Instructions

1. Compute per-sample depth/coverage with CoverM v0.7.0+ (or BBMap for short reads, minimap2 for long reads).
2. Bin contigs with **QuickBin** through Bryce Foster's official BBTools container (`bryce911/bbtools:39.85`; record digest when pulled). QuickBin is high-fidelity, CheckM2-agnostic, and scales well on both short-read and long-read assemblies. On a GPU node, run **SemiBin2 v2.3.0+** instead — self-supervised contrastive learning with CUDA-backed PyTorch. MetaBAT2 v2.18+ is kept only as a legacy fallback for reproducing prior pipelines.
3. Run `/tracking-taxonomy-updates` for BBTools-container QuickClade domain triage on the bin directory and the source assembly with `percontig`. Persist the per-contig screen so mixed bins are visible.
4. Route bins by the QuickClade domain screen:
   - Bacteria or Archaea -> run GTDB-Tk taxonomy assignment. If the GTDB-Tk reference package is missing, set it up under `$BIO_DB_ROOT`, export `GTDBTK_DATA_PATH`, run `gtdbtk check_install`, and record the release before classification.
   - Eukaryota -> run EukCC v2.1.3+ for eukaryotic bins.
   - Viral or virus-like -> remove from MAG QC and route candidate contigs/genomes to `/bio-viromics`; use vConTACT3 for phage/prokaryotic-virus evidence and GVClass for giant-virus/Nucleocytoviricota candidates.
   - Mixed or low-confidence -> flag as potential chimeras and inspect per-contig assignments before QC scoring.
5. Run domain-specific QC:
   - CheckM2 v1.1.0+ for bacterial and archaeal bins (note: v1.1.0 is a breaking upgrade — new DIAMOND v3 database from Zenodo DOI 10.5281/zenodo.14897628 and new dependency tree; re-install via mamba and refresh the DB).
   - EukCC v2.1.3+ for eukaryotic bins.
   - GUNC v1.0.6+ for contamination detection across all non-viral bins; treat it as a complement to CheckM2 (improves recall of chimeric bins).

## Quick Reference

| Task | Action |
|------|--------|
| Run workflow | Follow the steps in this skill and capture outputs. |
| Validate inputs | Confirm required inputs and reference data exist. |
| Review outputs | Inspect reports and QC gates before proceeding. |
| Tool docs | See `docs/README.md`. |

## Input Requirements

Prerequisites:
- Tools available in the active environment (Pixi/conda/system). See `docs/README.md` for expected tools.
- Reference DB root: set `BIO_DB_ROOT` to the project or site-local database directory.
- Coverage/depth tables or reads available to compute coverage.
- Docker or Apptainer/Singularity available for `bryce911/bbtools` QuickBin runs, or a documented local BBTools install.
Inputs:
- contigs.fasta
- coverage.tsv (per-sample depth table)

## Output

- results/bio-binning-qc/bins/
- results/bio-binning-qc/quickclade_percontig.tsv
- results/bio-binning-qc/domain_routing.tsv
- results/bio-binning-qc/gtdbtk_taxonomy.tsv
- results/bio-binning-qc/bin_metrics.tsv
- results/bio-binning-qc/bin_qc_report.html
- results/bio-binning-qc/logs/

## Quality Gates

- [ ] Completeness and contamination meet project thresholds.
- [ ] Chimera and contamination flags are below thresholds.
- [ ] On failure: retry with alternative parameters; if still failing, record in report and exit non-zero.
- [ ] Verify contigs.fasta and coverage.tsv are non-empty.
- [ ] Verify reference DBs for QC tools exist under the reference root.
- [ ] QuickClade `percontig` screen exists for the source assembly and bin set before CheckM2/EukCC/GTDB-Tk decisions.
- [ ] Bacterial and archaeal bins have GTDB-Tk taxonomy with the database release recorded.
- [ ] Viral/virus-like bins are routed to `/bio-viromics` instead of reported as MAGs.
- [ ] Mixed-domain bins are flagged as possible contamination/chimeras with per-contig evidence.

## Examples

### Example 1: Expected input layout

```text
contigs.fasta
coverage.tsv (per-sample depth table)
```

## Troubleshooting

**Issue**: Missing inputs or reference databases
**Solution**: Verify paths and permissions before running the workflow.

**Issue**: Low-quality results or failed QC gates
**Solution**: Review reports, adjust parameters, and re-run the affected step.

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [fmschulz](https://github.com/fmschulz)
- **Source:** [fmschulz/omics-skills](https://github.com/fmschulz/omics-skills)
- **License:** MIT
- **Homepage:** https://fmschulz.github.io/omics-skills/

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-fmschulz-omics-skills-bio-binning-qc
- Seller: https://agentstack.voostack.com/s/fmschulz
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.