Test for the comparability of the significance scores, for the 4 tested significance scoring methods, between fixed position motifs of different length and datasets of different size. (a) Boxplot comparisons of the 4 tested significance scoring schemes, for each dataset size and motif length, of the distribution of top ranking motifs significance values. The first boxplot in each panel is the uniform distribution. (b) The heatmap describes the Mann-Whitney p-value for all-by-all comparisons of the 3 dataset sizes and 3 motifs lengths. The Mann-Whitney, in this case, tests for rejection of the hypothesis that the distributions of Sig values of top ranking motifs of different length and dataset size are sampled from the same distribution.