GWAS Functional Annotation

A compact demo for GWAS QC and visualization, plus a scaffold for PolyFun-style annotation steps. The demo simulates 5,000 variants with seeded signals on chromosomes 1, 6, and 12, then writes a Manhattan plot you can sanity-check quickly. This repo is about workflow shape and data contracts, not study results.

Why it matters

• QC rules (chromosomes, p-values) usually break pipelines before modeling does.
• A small, reproducible Manhattan plot is a fast smoke test for GWAS data.
• PolyFun requires strict file layout; the scaffold makes those assumptions explicit.

Quickstart

make setup
make demo
make test

What you get

• reports/manhattan_demo.png (synthetic Manhattan plot)
• data/gwas_demo.csv (synthetic summary stats: chrom, pos, pval)
• reports/polyfun/polyfun_steps.json (if you run python src/polyfun_workflow.py)

Notes / assumptions

• QC in src/ingest_gwas.py drops rows with p-values outside [0,1] or chromosomes outside 1–22.
• The Manhattan demo uses the 5e-8 threshold line to mimic common GWAS significance.

Status Ready for demo; real annotations are intentionally out of scope.