BioProspector & MDScan: Discovering transcription factor binding sites in regulatory regions of co-regulated genes

TStanford大学のBrutlag博士らのグループで開発され，DoubleTwist社から提供されているMDScanとBioProspectorという配列上のプロモータ領域の配列解析ソフトウエアについての講演です。Brutlag博士からいただいた要旨もあわせてご案内いたします。

Gene expression experiments performed with whole genome microarrays can often give systematic analyses of sets of genes that are co-regulated under different experimental conditions. One would like to discover the regulatory basis for this co-expression. Is co-regulation due to a common promoter or due to common transcriptional factor binding sites in the upstream regulatory regions? The first method we developed, called BioProspector, can discover conserved DNA sequence motifs in upstream sequences of co-regulated genes taking into account all of the issues mentioned above. BioProspector uses a Metropolis algorithm known as Gibb's sampling to discover such consensus sequences. BioProspector can discover shorter and less well conserved motifs than other methods and in a sorter amount of time. We have also developed a new method called MDScan (Motif Discovery Scan) which, unlike BioProspector, is exhaustive and is 5-30 times faster than BioProspector or other commonly used methods. MDScan can also discover shorter and weaker conserved sequences than BioProspector or other programs.