Fig. 3

FDR for the synthetic study. The synthetic data was generated using noise variances σ²=[0.05,0.1,0.2,0.3,0.4]. Vertical axis is achieved FDR and horizontal axis the significance level used. Plot (a) gives the FDR for noise level σ²=0.05 for different number of bootstrap samples and the Oscope method for different cut-offs. Plot (b) compares the effect of different noise levels on the FDR of the bootstrap hypothesis test using 2000 samples