Briefings in Bioinformatics Advance Access originally published online on June 4, 2009
Briefings in Bioinformatics 2009 10(5):509-524; doi:10.1093/bib/bbp025
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Computational methods for the detection of cis-regulatory modules
Corresponding author. Peter Van Loo, Department of Human Genetics, VIB and University of Leuven, Herestraat 49, Box 602, B-3000 Leuven, Belgium. Tel: +32 16 33 01 44; Fax: +32 16 34 71 66; E-mail: peter.vanloo{at}med.kuleuven.be
Metazoan transcription regulation occurs through the concerted action of multiple transcription factors that bind co-operatively to cis-regulatory modules (CRMs). The annotation of these key regulators of transcription is lagging far behind the annotation of the transcriptome itself. Here, we give an overview of existing computational methods to detect these CRMs in metazoan genomes. We subdivide these methods into three classes: CRM scanners screen sequences for CRMs based on predefined models that often consist of multiple position weight matrices (PWMs). CRM builders construct models of similar CRMs controlling a set of co-regulated or co-expressed genes. CRM genome screeners screen sequences or complete genomes for CRMs as homotypic or heterotypic clusters of binding sites for any combination of transcription factors. We believe that CRM scanners are currently the most advanced methods, although their applicability is limited. Finally, we argue that CRM builders that make use of PWM libraries will benefit greatly from future advances and will prove to be most instrumental for the annotation of regulatory regions in metazoan genomes.
Keywords: transcription regulation, cis-regulatory modules, genome annotation, regulatory regions, computational CRM detection
Submitted: March 2, 2009. Received (in revised form): April 19, 2009.