Identifying splicing regulatory elements with de Bruijn graphs.

Area of Science:

• Genomics
• Molecular Biology
• Bioinformatics

Background:

• Splicing regulatory elements (SREs) are crucial for pre-mRNA splicing, influencing gene expression.
• Current methods for SRE identification are primarily experimental or limited by fixed-length computational approaches.
• Understanding SREs is vital for deciphering gene regulation and associated diseases.

Purpose of the Study:

• To develop a novel computational method for identifying variable-length SREs within exons.
• To combine genomic structure, word count analysis, and experimental data for SRE prediction.
• To provide a more comprehensive tool for discovering potential regulatory elements in the human genome.

Main Methods:

• Utilized de Bruijn graphs to model genomic sequences.
• Integrated word count enrichment analysis to identify significant sequence patterns.
• Incorporated experimental evidence to validate predicted SREs.
• Developed a method to identify SREs of variable lengths (6-15 nucleotides).

Main Results:

• Identified 2001 putative exonic enhancers and 3080 putative exonic silencers in human genes.
• The predicted SREs exhibit variable lengths, overcoming limitations of fixed-kmer approaches.
• A significant overlap was observed between predicted SREs and experimentally verified binding sites.
• The developed model offers a novel computational approach for SRE discovery.

Conclusions:

• The de Bruijn graph-based formalism provides an effective method for identifying variable-length SREs.
• This computational approach enhances the discovery of exonic enhancers and silencers.
• The findings facilitate further experimental validation and deepen the understanding of gene splicing regulation.