RAP: accurate and fast motif finding based on protein-binding microarray data.

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Protein-binding microarrays (PBMs) are high-throughput tools measuring transcription factor (TF) DNA binding.
• Existing algorithms for motif discovery from PBM data often produce motifs that are accurate for in vitro binding prediction or similar to literature motifs, but rarely both.

Purpose of the Study:

• To develop a novel, simple algorithm for inferring transcription factor binding-site motifs from PBM data.
• To improve upon the accuracy of in vitro binding prediction and the similarity to literature motifs compared to existing methods.

Main Methods:

• Development of a new algorithm for motif inference from PBM data.
• Evaluation of motif prediction accuracy for in vitro binding.
• Comparison of generated motifs with previously published literature motifs.
• Assessment of the impact of motif length and flanking sequence positions.

Main Results:

• The new algorithm outperforms prior art in both predicting in vitro binding and generating motifs similar to literature motifs.
• Results challenge the notion that lower information content motifs are better models for TF binding specificity.
• Side positions flanking the core motif significantly impact binding and should not be removed.
• A substantial decrease in prediction quality is observed when transitioning from in vitro to in vivo binding prediction across all methods.

Conclusions:

• The developed algorithm offers a superior approach to transcription factor motif discovery from PBM data.
• Motif length and flanking sequences are critical components for accurate binding site modeling.
• Further research is needed to bridge the gap between in vitro and in vivo binding prediction accuracy.