Figure 2 presents the power of the three trend tests of association at a 5% significance level for a high-risk allele frequency of 20%, as a function of the allelic odds ratio. Power is estimated over 5,000 replicates of data, with no divergence between the case-control population and the external cohorts (FST = 0). As expected, both tests that make use of the expanded control cohort are noticeably more powerful than T_CC. However, more importantly, there is no difference in power between T_F and T Fmds. Thus, by adjusting only for those axes of genetic variation that are correlated with phenotype, we take account only of population structure when it exists, and thus do not penalize T_Fmds. The same conclusions are reached, irrespective of high-risk allele frequency (see supplementary Fig. 1 for a high-risk allele frequency of 5%).
