Figure 1.
Power comparisons of seven tests when 50% causal variants have a positive effect on phenotype while the remaining 50% have a negative effect. The left panel considers 10% high/low extreme phenotype sampling with the three rows corresponding to 40%, 60%, and 80% causal variants. The right panel considers 20% high/low extreme phenotype sampling. Three sample sizes are considered: n = 500, 1000, 2000. Powers are estimated at the 0.05 significance level.