Population specific variant imputation accuracy. (A) Imputation accuracy for the variants on four superpopulations using HRC as the reference and 1000 Genomes as the query panel. Each box plot shows the genotype R2 distribution for the variants most frequently observed on the respective population. (B) The mean R2 difference between plaintext imputation and the unphased protocol at distinct allele frequency bins with HRC reference and 1000 Genomes as the query panel. Each bar corresponds to a population-specific variant set. (C) Imputation accuracy for population-specific variants using 1000 Genomes as the reference panel and GTEx as the query panel. (D) Mean genotype R2 difference between plaintext protocol and unphased protocol with 1000 Genomes as reference and GTEx panel as query.
