Assume that there are E samples with N genes each, and assume that in each sample, M genes are changed (either up- or down-regulated). What is the probability of obtaining at least K genes that are consistently misregulated throughout the E samples? This probability is given by Eq. (1) in the paper of Wahde et al. (submitted to Bioinformatics), which, in turn, uses Eqs. (3) and (4). An implementation (in Mathematica format) of Eqs. (3) and (4) can be downloaded here.