Fig. 7
From: Ancestry effects on type 2 diabetes genetic risk inference in Hispanic/Latino populations
Correcting for ancestry bias in T2D risk inference. The null distribution of T2D relative risk between the EA-MA populations is generated by randomly sampling risk alleles with the same frequency of ancestral/derived alleles as the GWAS source population. The initial observed T2D relative risk is compared to the null distribution to yield a corrected score. The observed T2 relative risk is shown on the distribution (arrow) along with the corrected T2D relative risk value and its significance (upper left)