Figure 2.

QQ-plots of genome-wide loci of the mouse selection experiment using four different methods. In each qq-plot, the y-axis is the observed test statistic and the x-axis is the expected test statistic under the null model. The upper two panels are the mixed model approach under the allelic model (A) and the genotypic model (B). The lower two panels show the plots from the regularized F test (C) and the regression model (D). Both of the mixed model approaches show more data points deviating from the diagonal lines than the other approaches, thus indicating higher statistical power.