The DNA-binding specificities of transcription factors can be used to computationally predict cisregulatory modules (CRMs) that regulate gene expression 1 . However, the absence of specificity data for the majority of transcription factors limits the wide-spread implementation of this approach. We have developed a bacterial one-hybrid system that provides a simple and rapid method to determine the DNA-binding specificity of a transcription factor. Using this technology, we successfully determined the DNA-binding specificity of seven previously characterized transcription factors and one novel transcription factor, the Drosophila factor Odd-skipped. Regulatory targets of Odd-skipped were successfully predicted using this information, demonstrating that the data produced by the bacterial one-hybrid system is relevant to in vivo function.Several methods exist for determining the DNA-binding specificity of a transcription factor. SELEX is the most commonly employed method to define DNA-binding specificity 2 . More recently, microarrays of short oligonucleotides 3 or intergenic sequences 4 have been used to characterize transcription factors. Genome-wide chromatin immunoprecipitation (ChIPchip) 5 and DNA immunoprecipitation with microarray detection (DIP-chip) 6 have also been used in conjunction with computational analysis to identify statistically overrepresented sequence motifs to extract DNA-binding specificity from the genomic segments that are bound by a transcription factor. These methods, while powerful, have drawbacks: the in vitro technologies require the purification of the transcription factor in its active form; SELEX requires multiple rounds of selection to complete; and microarray-based techniques require the facilities and expertise to analyze the arrays and resulting data.We have developed a bacterial one-hybrid (B1H) system that provides a simple method for defining the DNA-binding specificity of a transcription factor. The selection procedure is rapid, because only a single round of selection is required to generate a set transcription factor binding sites, and it is readily accessible, because only basic molecular biology expertise is required to employ the technology. Conceptually, this system is similar to yeast one-hybrid systems that can determine the DNA-binding specificity of a transcription factor 7 and detect protein-DNA interactions 8 . However, a bacterial selection system provides advantages over the