The question addressed here is how cooperative interactions among transcription factors (TFs), a very frequent phenomenon in eukaryotic transcriptional regulation, can be used to identify genes that are regulated by one or more TFs with known DNA binding specificities. Cooperativity may be homotypic, involving binding of only one transcription factor to multiple sites in a gene's regulatory region. It may also be heterotypic, involving binding of more than one TF. Both types of cooperativity have in common that the binding sites for the respective TFs form tightly linked 'clusters', groups of binding sites often more closely associated than expected by chance alone. A statistical technique suitable for the identification of statistically significant homotypic or heterotypic TF binding site clusters in whole eukaryotic genomes is presented. It can be used to identify genes likely to be regulated by the TFs. Application of the technique is illustrated with two transcription factors involved in the cell cycle and mating control of the yeast Saccharomyces cerevisiae, indicating that the results obtained are biologically meaningful. This rapid and inexpensive computational method of generating hypotheses about gene regulation thus generates information that may be used to guide subsequent costly and laborious experimental approaches, and that may aid in the assignment of biological functions to putative open reading frames.