Box plot of the CV of p-value histogram under different correlation structures. (a) datasets having 100 genes, with 1 or 10 genes per block. (b) datasets having 1000 genes, with 1, 10 or 100 genes per block. (c) datasets having 10000 genes, with 1, 10, 100 or 1000 genes per block. The histogram was calculated by splitting p-values into 20 bins between [0, 1] with the width of each bin being 0.05. The horizontal lines represent the expected CV when genes are independent.