Disease case study

Skill Alim430/bioresearch-agent/skills/biomedical/disease-case-study

Executable biomedical workflows for AI assistants — literature analysis, biomarker discovery, and Mendelian randomization through reproducible agent skills.

Install

npx -y skills add Alim430/bioresearch-agent --skill disease-case-study

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Run a reproducible, real-public-data disease case study that composes the literature → biomarker → causal workflows end-to-end and emits a blind benchmark evaluation (known-gene recovery, pathway sanity, reproducibility hash). Use when the user wants to prove the framework on a real disease (e.g. Parkinson's GSE7621) or asks for a "case study", "validation run", or "benchmark" of the biomedical workflows.

SKILL.md

2.3 KB, as published. Nobody here has run it

BioResearch Agent — Disease Case Study Skill

Capability

A workflow-composition + validation skill. It runs a real, public GEO dataset through the framework's analysis engine (differential expression → pathway enrichment → candidate ranking, with optional literature and causal-inference stages) and produces a blind benchmark evaluation:

Known-gene recovery — how many established disease genes surface in the top-ranked candidates (e.g. SNCA / LRRK2 / PARK7 / PINK1 / PRKN / GBA for Parkinson's).
Pathway sanity — whether disease-relevant pathways (dopamine / mitochondrial / oxidative / immune) are enriched.
Reproducibility — commit hash, environment, and data sha256 recorded in an Evidence Package.

Case Study 1 ships with the suite: Parkinson's disease / GSE7621 (GPL570, 25 samples). See bio-research-os/eval/case_study_pd.py and the Biomedical Workflow Validation Suite README.

Run

python bio-research-os/eval/case_study_pd.py \
    --matrix-path downloads/gse7621_matrix.txt.gz \
    --annot-path  downloads/gpl570.annot.gz \
    --output-dir  docs/case-study

(If paths are omitted the runner downloads the real GSE7621 matrix + GPL570 annotation itself.)

Outputs (in `docs/case-study/`)

GSE7621_deg.csv — differential expression results
GSE7621_top_candidates.csv — ranked biomarker candidates
GSE7621_pathway_enrichment.csv — enriched KEGG / GO terms
GSE7621_volcano.png — volcano plot
GSE7621_report.txt — structured report incl. blind benchmark
GSE7621_evidence_package.json — provenance + benchmark + limitations

Note

This skill is a composition & validation interface. It adds no statistics of its own — every number comes from the framework's workflow modules. It is the honest way to demonstrate the framework works on real data (not synthetic injection). Requires no LLM key.