TABLE 1.

Estimated power when n_G G-only effects are present (n_G ∈ {10, 20, 40, 80}) and the total amount of variation explained is fixed at 40%$\left(R_G^2=0.4/n_G\right)$. One-step GWIS: The standard one-step GWIS. RB-Weighted: Rank-based weighted hypothesis testing proposed by Ionita-Laza et al. (2007) with B₀ = 5; SB-Weighted: Our proposed significance-based weighted hypothesis testing with τ = (0, 5∕25000, 15∕25000, … , 1) as the p-value cutoffs. Bonferroni: Standard Bonferroni correction within bin; simpleM: The simpleM procedure proposed by Gao et al. (2008) with C = 0.995. Results are averaged over 5000 simulations.

nG =	10	20	40	80
$R_G^2=$	0.04	0.02	0.01	0.005
One-step GWIS	0.439	0.437	0.443	0.442
RB-Weighted
Bonferroni	0.694	0.670	0.726	0.801
simpleM	0.713	0.688	0.736	0.812
SB-Weighted
Bonferroni	0.814	0.797	0.796	0.812
simpleM	0.829	0.811	0.812	0.826