highly significantP = 4×10 −4
The difference in the two distributions is highly significant (Fisher's exact test, P = 4×10 −4 ), suggesting that ethnic differences contribute to the bimodal distribution of imputation quality seen in Figure S12 .
The difference in the two distributions is highly significant (Fisher's exact test, P = 4×10 −4 ), suggesting that ethnic differences contribute to the bimodal distribution of imputation quality seen in Figure S12 .
Table S8 shows that only the Immunochip loci showed a nominally significant ( P = 0.001) excess of high-ranking BFs (those in the top 5% of the empirical distribution) of genes.