Estimated density distribution $\hat{D}(p)$ of p-values for single-window objects (green), and for multiple-window objects with p-values combined using the Fisher method (magenta) and combined with the min(p) method (blue). Top panel: p-value distributions from the FPR data. The inset shows a zoom of the area delimited by the dashed line. Bottom panel: Results from one noise-only simulation run.