Summary of estimates of the ancestry of individuals in the population with recent admixture (data set B). The peaks, from left to right, represent histograms of estimated values of q(i)1 (the proportion of ancestry of individual i in subpopulation 1) for individuals with 0, 1, 2, 3, and 4 grandparents, respectively, in subpopulation 1. The solid lines show the results for the entire sample, while the dotted lines show the results for those individuals who are affected. These data were simulated under a model in which an individual’s risk of disease increases with their amount of ancestry in subpopulation 1, with the result that affected individuals tend to have a greater proportion of ancestry in subpopulation 1 than do controls. This effect can be seen here, as the proportion of affected individuals in each peak increases from left to right. Our test for association uses the estimated q(i) to control for the presence of biased sampling.