Skip to main content

Table 1 Pearson correlation between Nei’s G ST and imputation quality scores of the 20 POPRES populations in dependence on reference panel, imputation software and measure of imputation quality, CIG = correctly imputed genotypes

From: Impact of genetic similarity on imputation accuracy

Reference Software HELLI > =0.45 (%) SEN > =0.45 (%) CIG (%) Rsq/Info mean Rsq/Info > =0.8 (%)
CEU MaCH −0.907 −0.911 −0.902 −0.959 −0.930
MaCH_minimac −0.870 −0.871 −0.868 −0.914 −0.865
IMPUTE2 −0.877 −0.888 −0.876 −0.943 −0.898
YRI MaCH −0.962 −0.969 −0.966 −0.975 −0.849
MaCH_minimac −0.950 −0.961 −0.953 −0.971 −0.620
IMPUTE2 −0.953 −0.966 −0.956 −0.969 −0.507
CHB. JPT MaCH −0.907 −0.905 −0.910 −0.767 −0.739
MaCH_minimac −0.892 −0.893 −0.896 −0.767 −0.785
IMPUTE2 −0.853 −0.852 −0.856 −0.784 −0.806
MEX MaCH −0.917 −0.915 −0.914 −0.946 −0.923
MaCH_minimac −0.898 −0.892 −0.898 −0.908 −0.882
IMPUTE2 −0.898 −0.902 −0.901 −0.932 −0.904
  1. By estimating the association between G ST and imputation accuracy score and by computing a test of the correlation being zero, we got p-value < 9.52e-13 for all three scenarios