Identifying cis-regulatory sequences by word profile similarity

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Identifying regulatory sequences in genomes remains a significant challenge.
• Despite abundant genomic data and validated examples, efficient recognition is difficult.

Purpose of the Study:

• To develop a simple, motif-agnostic approach for searching regulatory sequences.
• To identify co-regulated genomic regions based on compositional similarity.

Main Methods:

• The method searches for regulatory sequences using compositional similarity to known cis-regulatory elements.
• It leverages sequence motifs within known regulatory sequences to find co-regulated regions.
• Local sequence dissimilarity is explored as a universal feature for regulatory sequence identification.

Main Results:

• The approach successfully recovers sequences under similar regulatory control without predefined motifs.
• Shared words in recovered sequences often indicate known binding sites.
• Local dissimilarity proves to be a more universal feature for regulatory sequences in Drosophila than local word profile clustering.

Conclusions:

• The developed method effectively identifies co-regulated sequences by utilizing sequence motifs and compositional analysis.
• Local sequence dissimilarity is a novel and effective feature for distinguishing regulatory sequences genome-wide.
• Source code for the WPH-finder tool is publicly available.