Validation datasets, code, and plots for the Archipelago method.

Archipelago Validation This directory contains validation studies for the Archipelago method, a framework for integrating single-variant GWAS and variant set association test (VSAT) statistics. . ├── README.md ├── data │ ├── 1kg │ ├── ukbb │ └── ukbb_phewas ├── output │ ├── archipelago_plot_1kgeas.png │ ├── archipelago_plot_1kgeas_enriched_variant_frequencies.png │ ├── archipelago_plot_ukbb_PheWAS.png │ ├── archipelago_plot_ukbb_full.png │ ├── archipelago_plot_ukbb_subset.png │ ├── archipelago_plot_vsat_raw_1kgeas.png │ └── archipelago_plot_vsat_raw_ukbb.png └── src ├── 1kg_archipelago.R ├── 1kg_prep ├── ukbb_archipelago_v0.R └── ukbb_phewas_archipelago_v0.R Input The manuscript included three validation datasets: data/1kg/ — 1000 Genomes Project (East Asian samples) read GWAS trait and simulated pathway-level trait. data/panukbb/ — Pan-UK Biobank (platelet distribution width, DeepRVAT) real GWAS and gene-level trait. data/ukbb_wgs/ — UK Biobank WGS (UTR collapsing PheWAS) real GWAS and gene-level trait. Each data subdirectory contains its own README with workflow details, source data, and outputs used in the validation studies. For large datasets like raw genotypes in 1KG or large UKBB summary statistics, we provide the direct download link to public the source. The UKBB studies are fully self contained in the R scripts. The 1KG study additionally does the full GWAS prep in the 1kg_prep directory. Output The output/ folder contains final plots from the three validation studies: 1KG - Archipelago Plot: archipelago_plot_1kgeas.png - VSAT Raw Plot: archipelago_plot_vsat_raw_1kgeas.png Pan-UKBB - Archipelago Plot: archipelago_plot_ukbb_full.png - VSAT Raw Plot: archipelago_plot_vsat_raw_ukbb.png UKBB WGS - Archipelago Plot: archipelago_plot_ukbb_wgs.png - VSAT Raw Plot: archipelago_plot_vsat_raw_ukbb_wgs.png Additional outputs: - archipelago_plot_1kgeas_enriched_variant_frequencies.png — enriched variant frequency summary for 1KG- archipelago_plot_ukbb_subset.png — downsampled version of the UKBB plot for display Scripts Reproducible plotting scripts are available under `src/`: `1kg_archipelago.R` `panukbb_archipelago_v0.R` `ukbb_wgs_archipelago_v0.R` Additionally, to provide a complete workflow for 1kg we include `src/1kg_prep/` which has the full variant-level GWAS and pathway-level VSAT. The readme provides the tutorial source which matches its directory layout. These scripts prepare the input data and run the final integrated Archipelago visualisations per study Citation For methodology and results, see: Archipelago method for variant set association test statisticsDylan Lawless, et al. medRxiv 2025.03.17.25324111 https://doi.org/10.1101/2025.03.17.25324111