CodingMotif: exact determination of overrepresented nucleotide motifs in coding sequences

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Coding DNA sequences contain regulatory motifs beyond protein-encoding information.
• Detecting overrepresented motifs in coding regions is challenging due to protein-level constraints.
• Existing sampling and parametric methods have limitations in exploring sequence space and accuracy.

Purpose of the Study:

• To develop a novel algorithm for identifying nucleotide-level motifs with unusual copy numbers in protein-coding regions.
• To overcome the limitations of existing methods for motif detection in coding sequences.

Main Methods:

• Developed an O(N(log N)^2)-time algorithm named CodingMotif.
• Utilized a dynamic programming approach to calculate motif occurrence distributions across all coding sequences for a given amino acid sequence.
• Incorporated background models for codon usage and dinucleotide biases.
• Leveraged the sparseness of motif loci to accelerate convolution calculations.

Main Results:

• CodingMotif accurately assesses the non-parametric p-value for motif over- or under-representation.
• The algorithm demonstrates superior accuracy in identifying known functional motifs compared to sampling and parametric methods.
• Successfully applied to various coding datasets, including ChIP-seq data for NRSF and GABP transcription factors.

Conclusions:

• CodingMotif represents a significant advancement in detecting overrepresented motifs within coding sequences.
• The tool is expected to aid in identifying motifs in functional genomic datasets, including DNA-protein, RNA-protein, and microRNA-RNA binding interactions.
• A software implementation is publicly available for use.