Biomedical FigureYa Atlas

Biomedical FigureYa Atlas is a Claude/Codex skill for biomedical scientific figure generation, visual reference matching, and R plotting workflow reuse.

It extends the public FigureYa project into a curated figure atlas for AI-assisted research plotting. The skill combines a searchable catalog, visual preview gallery, categorized biostatistics and bioinformatics indexes, and lightweight R/Rmd reference code.

What This Skill Provides

• A biomedical scientific figure gallery and visual catalog
• 306 manually reviewed preview images and 45 montage pages
• 312 FigureYa-derived plotting tool records
• Separate indexes for general biostatistics and bioinformatics/omics figures
• Reference R Markdown and helper R scripts for adapting plot workflows
• GitHub fallback links for each FigureYa module
• Color palette guidance for publication-style figures

Use Cases

Use this skill when you need to:

• Create R code for biomedical or scientific plots
• Match a manuscript figure style to a reference implementation
• Choose between similar plot types such as heatmaps, ROC curves, survival curves, volcano plots, enrichment plots, single-cell figures, mutation/CNV plots, and multi-omics figures
• Convert FigureYa-style examples into reusable plotting templates
• Find paper-quality color and layout patterns for biomedical visualization

Structure

biomedical-figureya-atlas/
├── SKILL.md
├── AGENTS.md
├── CLAUDE.md
├── agents/
│   └── openai.yaml
├── assets/
│   ├── previews/
│   └── montages/
├── references/
│   ├── figure-contract.md
│   ├── retrieval-workflow.md
│   ├── visual-matching.md
│   ├── r-code-adaptation.md
│   ├── qa-contract.md
│   ├── figureya_catalog.json
│   ├── figureya_index.md
│   ├── index_biostatistics.md
│   ├── index_bioinformatics.md
│   ├── visual_catalog.json
│   ├── visual_index.md
│   ├── visual_exclusions.json
│   ├── color_schemes.md
│   └── code/
└── scripts/

How AI Agents Should Use It

1. Read SKILL.md for the workflow.
2. Open references/figure-contract.md to classify task mode, domain, plot family, input shape, and output target.
3. Use references/retrieval-workflow.md plus references/index_biostatistics.md or references/index_bioinformatics.md to find candidates.
4. Use references/visual-matching.md, references/visual_index.md, and assets/montages/ when visual style matters.
5. Use references/figureya_catalog.json and references/visual_catalog.json for structured filtering and GitHub fallback links.
6. Read only the selected 1-3 modules under references/code/FigureYaXXX/ before adapting R code.
7. Use references/r-code-adaptation.md for script generation and references/qa-contract.md before final delivery.

Maintenance

The scripts/ directory contains utilities to rebuild the atlas from a local FigureYa checkout:

• scan_figureya.py rebuilds the machine-readable catalog.
• generate_index.py rebuilds Markdown indexes.
• create_montages.py syncs preview images from the FigureYa web gallery and rebuilds montage pages.
• sync_reference_code.py copies lightweight .Rmd, .rmd, and .R source files without copying large HTML reports or data files.

HTML reports are intentionally not bundled because they are large generated outputs. Use web_reports in references/visual_catalog.json or the upstream GitHub links when full reports are needed.

Attribution

This skill is derived from the public FigureYa repository:

https://github.com/ying-ge/FigureYa

Please acknowledge FigureYa when using its plot patterns, code structure, gallery images, or examples. Follow the citation and license information from the upstream FigureYa project.

License

This repository is released under CC BY-NC-SA 4.0 to align with the upstream FigureYa licensing terms. This means attribution is required, commercial use is restricted, and derivative works should be shared under the same license.

Third-party packages, libraries, datasets, and external tools referenced by the bundled R/Rmd examples retain their own licenses and citation requirements.

Safety

Treat bundled R/Rmd files as reference implementations. Review code before execution, especially install scripts, downloaded dependencies, network calls, and file-system operations. Do not upload private research data to external services. Check package licenses, data-source requirements, and upstream citation requirements before reusing code in a manuscript or public project.