Table 1 Imputation performance summary

From: A reference haplotype panel for genome-wide imputation of short tandem repeats

Panel (n = number of samples) Observed concordance Naive concordance Random concordance Observed length r2 Random length r2 Observed allelic r2 Random allelic r2
SSC—LOO (n = 1916) 96.7% 71.7% 61.0% 0.906 0.605 0.861 0.552
SSC—LOO (multi-allelic) 94.3% 62.2% 48.5% 0.888 0.334 0.800 0.333
1000 Genomes—EUR (n = 49) 97.0% 75.1% 63.2% 0.921 0.678 0.892 0.543
1000 Genomes—EUR (multi-allelic) 94.8% 66.6% 50.0% 0.900 0.334 0.828 0.314
1000 Genomes—AFR (n = 46) 90.6% 70.2% 57.9% 0.746 0.619 0.706 0.493
1000 Genomes—AFR (multi-allelic) 85.6% 61.1% 44.4% 0.708 0.336 0.653 0.310
1000 Genomes—EAS (n = 45) 93.8% 77.2% 66.0% 0.823 0.690 0.781 0.557
1000 Genomes—EAS (multi-allelic) 89.4% 69.7% 53.7% 0.780 0.336 0.663 0.313
  1. Results indicate mean across all STRs analyzed. Allelic r2 values include all common alleles (frequency at least 5%). Multi-allelic refers to STRs with three or more common alleles. Naive and random denote the two null imputation models as defined in the Methods