Figure 1

Analysis flow. The input consists of a set of coexpressed or coregulated genes. Step 1: Screening the intergenic sequences of these genes with a library of PWMs. Step 2: Apply our itemset mining strategy to find all the modules (closed motifsets) that occur in a minimal number of genes in the dataset (a minimum support defined by the user). Step 3: Determine the final rank of each module. An original ranking is assigned to each module based on the module specificity score. In the update step the rank of overlapping modules is reduced by iteratively assigning an updated score.