Computational inference of transcription factor cooperation by fuzzy frequent itemset mining

Eucaryotic gene control regions consist of a promoter plus regulatory DNA sequences which may appear distant from the gene promoter. Regulatory proteins (called transcription factors, TFs), coordinately bind to these regions (TF binding sites, TFBSs) and produce the correct gene expression patterns. We present a novel fuzzy approach to study significant co-occurences of closely located TFBSs in the yeast whole-genome. Our approach takes advantage of the ability of fuzzy techniques to handle imprecision, inherent to regulatory-regions location data. The methodology is based on a fuzzy frequent itemset mining algorithm and overcomes some of the limitations of previous approaches. The results obtained from the yeast genome allow us to propose the application of the procedure over more complex genomes in future works.