Table 1.
Number of classical HLA alleles and polymorphic amino acid positions in the HLA reference panel.
Table 2.
Concordance rates between imputed and genotyped alleles.
Figure 1.
Concordance rates using various reference panels.
SNPs present in 9 different commercial arrays were extracted for imputing the HLA alleles from the HapMap3 r2 dataset of East Asian (CHB + JPT) individuals. Each subset was imputed for HLA-A, -B, -C, -DQB1 and -DRB1 using the Korean reference panel. Concordance rates (y-axis) were plotted against the proportion of reference-panel SNPs that were present in each subset (x-axis).
Figure 2.
Frequencies and disease effect sizes of imputed and genotyped HLA-DRB1 alleles.
We revisited our previous rheumatoid arthritis association studies using either typed SNPs [16] or HLA alleles [17]. After imputing HLA variants from the SNP-based dataset, (A) frequencies and (B) disease effect sizes of the imputed classical alleles of HLA-DRB1 were compared with those of genotyped classical alleles in the HLA-based dataset.