Reference code for the manuscript "An Expanded Reference for Middle Eastern Populations Improves GWAS" prediction.
harmonization/
├── harmonization.md Description of code used to generate data for figures 1 and 2:
│ Harmonize multiple MID cohorts into one dataset:
│ liftover, QC, merge, downsampling, PCA, ADMIXTURE, Fst,
│ and clinically relevant SNP annotation
├── clinical_snps_regions.R Intersect region-level allele
│ frequencies with ClinPGx/GWAS Catalog annotations
├── Fig01_target_pca_map.R Code used to generate Figure 1
├── Fig02_neighboring_pca_admix.R Code used to generate Figure 2
├── pca_country_ethnicity_study.R Code used to generate Figure S3
└── plot_fst_heatMap.R Code used to generate Figure S7
ancestry_inference/
├── reference_qc.md Description of code used to build the
│ expanded reference panel (HGDP-1kGP-Almarri):
│ per-cohort QC, merge, post-merge QC, liftover to hg37
├── plink_ukbb/ Code used for genetic ancestry inference and GWAS in UK Biobank:
│ Intersect UKBB with the reference panel, train a
│ random forest ancestry classifier, predict and refine
│ MID ancestry calls, and validate against place-of-birth
│ and GWAS
└── hail_aou/ Code used for genetic ancestry inference and GWAS in All of us
visualization/
├── HoneycombPCA.R PCA clustering and Honeycomb plot function
└── LabelFreeADMIX.R ADMIXTURE ordering and plotting function