Bio Methods Writing

Bioinformatics Methods Writing

Use this skill to turn workflow provenance into a reproducible Methods section. It documents what was done; it does not execute or repair the workflow.

Boundary

This skill does:

• draft Methods subsections for data sources, preprocessing, QC, statistics, omics analyses, validation, and software availability
• normalize workflow notes into journal-ready prose
• identify missing reproducibility fields
• keep methods aligned with evidence packs and analysis logs

This skill does not:

• run pipelines, debug code, validate accession records, or choose parameters after the fact
• fabricate package versions, genome builds, reference databases, thresholds, random seeds, sample exclusions, or command options
• write Results interpretations

If provenance is too incomplete, produce a Methods skeleton plus missing-fields table and route analysis/provenance recovery to bioinfo-evidence-orchestrator or reproduction skills.

When to open extra files

Workflow

1. Collect provenance. Use workflow tables, commands, notebooks, logs, README files, config files, and EVIDENCE_PACK.md.
2. Group by reproducible step. Typical order: data acquisition, preprocessing, QC, normalization, analysis, statistics, validation, visualization, software.
3. Write methods with enough detail to repeat. Include versions, parameters, reference assets, thresholds, and exclusion criteria only when supplied.
4. Separate methods from results. Do not include observed outcomes unless needed to define filtering or QC criteria.
5. Normalize terminology. Use standard names for repositories, assays, packages, genome assemblies, annotations, and statistical models.
6. Create a missing-fields table. Mark what the author must confirm.

Output format

Methods draft
[subsections]

Reproducibility table
| Step | Tool/package | Version | Key parameters | Input/output | Missing fields |
|---|---|---|---|---|---|

Author confirmation needed
- ...

中文核对
- ...