NUBIScan, an in silico approach for prediction of nuclear receptor response elements

Nuclear receptors (NRs) are transcription factors activated by a multitude of hormones, other endogenous substances, and exogenous molecules. These proteins modulate the regulation of target genes by contacting their promoter or enhancer sequences at specific recognition sites. The identification of these response elements is the first step toward detailed insight into the regulatory mechanisms affecting a gene. We have developed NUBIScan, a computer algorithm to predict DNA recognition sites for NRs in the regulatory regions of genes. The algorithm is based on weighted nucleotide distribution matrices and combines scores from both half-sites necessary for NR dimer binding. It provides more specific identification of functional sites than previous in silico approaches, as evidenced by scanning published regulatory regions of drug-inducible genes and comparing the obtained predictions with experimental results. In prospective analyses, NUBIScan consistently identified new functional NR binding sites in sets of large sequences, which had eluded previous analyses. This is exemplified by the detailed functional analysis of the flanking region of two genes. This approach therefore facilitates the selection of likely sites of gene regulation for subsequent experimental analysis.