Figure 5

Schematic showing the method employed in counting the co-occurrence of the predicted motifs. Each DNA sequence was divided into N _b bins of 30,000/N _b (=L _b ) bp (left; N _b and L _b are set to 30 and 100, respectively, in this figure). Co-occurrences for each bin-set are identified for each motif set (motifs A and B; sets enclosed in rectangles in the left-hand figure). For example, the count of the co-occurrence of the motif set (A, B) on promoter-2 was two; therefore, in the contingency table, the counts of the cells (A3, B5) and (A3, B6) were increased by one (right). Overlapping motif pairs were not counted. Based on the counts of the co-occurrences in all promoters, we constructed a contingency table for each motif pair.