Data harmonization and phasing: We then intersected and jointly phased the post-QC’ed cohort data with autosomal data from 247 1KGP reference panel individuals, removing conflicting sites and flipping any remaining strand flips. The merged dataset was then filtered to include only informative SNPs present in both the cohort and reference panel using a filter of MAF ≥ 0.05 and a genotype missingness cutoff of 90%. The program SHAPEIT285 was used to phase chromosomes, informed by the HapMap combined b37 recombination map86. Individuals from the cohort and reference panel were then separated and exported as harmonized sample and reference panel VCFs to be fed into RFMix87.
