Ancestry Analysis

Pipeline and lightweight report app for preparing 23andMe genotype data against reference panels, running EAGLE2, and publishing app-ready ancestry analysis outputs.

What This Repository Contains

• WSL genetics pipeline for 23andMe, 1000 Genomes, and HGDP data
• shared-SNP extraction and reference-aligned VCF generation
• HGDP general/local population labeling
• 1000 Genomes population labeling
• EAGLE2 v2.4.1 runner
• FLARE local ancestry runner
• RFMix local ancestry comparator
• holdout validation, model-tournament, and synthetic-admixture calibration helpers
• optional HMM smoothing for FLARE local ancestry segments
• static report UI with Report, Sample models, Validation, and Quality tabs

Main Workflow

From WSL:

cd /mnt/d/Python/Genetics
bash run_pipeline_wsl.sh

After shared-SNP extraction finishes, run EAGLE2:

cd /mnt/d/Python/Genetics
SKIP_EXTRACTION=1 RUN_EAGLE=1 THREADS=8 bash run_pipeline_wsl.sh

After EAGLE2, run a chr22 FLARE smoke test:

cd /mnt/d/Python/Genetics
SKIP_EXTRACTION=1 RUN_FLARE=1 FLARE_CHROMS=22 THREADS=4 FLARE_THREADS=4 FLARE_MEMORY_GB=6 bash run_pipeline_wsl.sh

Then run FLARE across all autosomes:

cd /mnt/d/Python/Genetics
SKIP_EXTRACTION=1 RUN_FLARE=1 FLARE_CHROMS=all THREADS=8 FLARE_THREADS=4 FLARE_MEMORY_GB=6 bash run_pipeline_wsl.sh

Run an unbiased HGDP holdout validation before trusting the model:

cd /mnt/d/Python/Genetics
SKIP_EXTRACTION=1 RUN_VALIDATE=1 VALIDATE_CHROMS=22 VALIDATION_SAMPLES_PER_LABEL=2 bash run_pipeline_wsl.sh

Run the first model tournament:

cd /mnt/d/Python/Genetics
SKIP_EXTRACTION=1 RUN_TOURNAMENT=1 TOURNAMENT_CHROMS=22 VALIDATION_SAMPLES_PER_LABEL=2 bash run_pipeline_wsl.sh

Run the independent RFMix chr22 comparator:

cd /mnt/d/Python/Genetics
SKIP_EXTRACTION=1 RUN_RFMIX=1 RFMIX_CHROMS=22 THREADS=4 FLARE_THREADS=4 bash run_pipeline_wsl.sh

Run RFMix validation and a small parameter check:

cd /mnt/d/Python/Genetics
SKIP_EXTRACTION=1 RUN_RFMIX_VALIDATE=1 RFMIX_VALIDATE_CHROMS=22 VALIDATION_SAMPLES_PER_LABEL=2 bash run_pipeline_wsl.sh
SKIP_EXTRACTION=1 RUN_RFMIX_GRID=1 RFMIX_VALIDATE_CHROMS=22 RFMIX_GRID_PRESETS=default,shorter_windows bash run_pipeline_wsl.sh

Create synthetic admixed targets for validation/calibration only:

cd /mnt/d/Python/Genetics
SKIP_EXTRACTION=1 RUN_SIMULATIONS=1 SIMULATION_CHROMS=22 bash run_pipeline_wsl.sh

Output Locations

Main generated outputs are written outside the repo:

/mnt/f/data/processed/genetics_eagle

App-facing JSON is also copied into:

data/

The chromosome view uses:

data/phasing_qc.json
data/chromosome_segments_hgdp.json
data/validation_hgdp_chr22.json
data/model_tournament_hgdp_chr22.json
data/validation_dashboard.json
data/sample_model_outputs.json

Documentation

Detailed pipeline notes:

README_pipeline.md

UI integration notes:

UI_INTEGRATION_BRIEF.md

HGDP label grouping audit:

docs/HGDP_LABEL_GROUPING.md

Lazaridis/EuropeFullyPublic modern-label audit and phasing decision:

docs/LAZARIDIS_LABEL_AUDIT.md
docs/LAZARIDIS_STAGE_DECISION.md

App

Run the static report app locally:

npm start

Then open:

http://localhost:4173