Identifying cis-regulatory modules by combining comparative and compositional analysis of DNA

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Predicting cis-regulatory modules (CRMs) in higher eukaryotes is computationally challenging.
• Existing methods for CRM prediction often rely on transcription factor binding site (TFBS) models, limiting their scope.
• There is a need for more general methods that do not require prior TFBS information for comprehensive CRM prediction.

Purpose of the Study:

• To develop a novel computational method for predicting CRMs.
• To identify CRMs using sequence conservation and statistical properties of DNA.
• To evaluate the performance of the new method against existing CRM prediction approaches.

Main Methods:

• Developed CisPlusFinder, a method that identifies perfect local ungapped sequences (PLUSs) indicative of core TFBS motifs.
• Utilized multiple species conservation to identify regions of high TFBS motif density.
• Assessed CisPlusFinder performance on a benchmark dataset of CRMs from early Drosophila development.

Main Results:

• CisPlusFinder predicted more annotated CRMs than all other tested methods on the benchmark dataset.
• Some previously identified 'false positive' predictions by CisPlusFinder corresponded to recently annotated CRMs in the REDfly database.
• The method demonstrates the effectiveness of combining comparative genomics with statistical DNA properties for CRM prediction.

Conclusions:

• CisPlusFinder offers a robust approach for CRM prediction, particularly valuable in the absence of known TFBS motifs.
• The study highlights the utility of multi-species conservation and statistical sequence properties for identifying regulatory elements.
• This method has the potential for genome-wide application in CRM discovery.