Figure 3

Quantil-quantil plot of the observed and expected –log(p-values) obtained by simulation. Reference distribution was an independent and uniform distribution U ∼ (0, 1) for 1018 p-values simulated (black dotted line). Test1(scenario1) = under dependent (LD) and standardization by $\mathrm{Var}\left(\widehat{\mathit{g}}\right)$ (blue dotted line). Test1(scenario2) = under independent (LE) and standardization by $\mathrm{Var}\left(\widehat{\mathit{g}}\right)$ (green dotted line). Test2(scenario2) = under independent (LE) and standardization by PEV (orange dotted line). Each scenario has 1018 p-values permuted 200 times. Bands represent confidence intervals of 95% (blue band = test1(scenario1), green band = test1(scenario2), pink band = test2(scenario2).