Comparison of six regulatory similarity measures. This figure is plotted as follows. First, a subset of one million samples was randomly selected from the evaluation dataset of 19495447 gene pairs. Second, the scores of the selected gene pairs are calculated. Third, the correlation plot of one subset is generated. Each point represents 2% of gene pairs (each figure contains 50 points) in that subset. In a correlation plot, x and y are the average RS score and the fraction of gene pairs that are co-regulated (validated by the literature evidence deposited in the YEASTRACT database) by at least one common TF, respectively. The gene pairs were sorted by the RS score. For example, the rightest point represents the 2% gene pairs of the highest RS scores. Finally, the three steps are repeated 100 times and this figure shows an average plot of 100 correlation plots. The dashed line indicates a random predictor in which RS scores are randomly assigned.