PairK: Pairwise k-mer alignment for quantifying protein motif conservation in disordered regions

Area of Science:

• Molecular Biology
• Bioinformatics
• Computational Biology

Background:

• Protein-protein interactions are fundamental to cellular processes.
• These interactions are often mediated by domains binding to short linear motifs (SLiMs) within disordered protein regions.
• Predicting these interactions is vital for understanding biological networks and formulating hypotheses, but current methods struggle with false positives.

Purpose of the Study:

• To develop a novel computational method for accurately predicting biologically relevant short linear motifs (SLiMs).
• To overcome limitations in quantifying sequence conservation within disordered protein regions, which hinders SLiM identification.
• To improve the prediction of functional domain-SLiM interactions for mapping protein networks.

Main Methods:

• Introduction of PairK (pairwise k-mer alignment), a novel, multiple sequence alignment (MSA)-free method.
• PairK quantifies motif conservation specifically within disordered protein regions.
• Comparison of PairK's performance against standard MSA-based conservation scores and a large language model (LLM)-based predictor.

Main Results:

• PairK significantly outperforms both MSA-based and LLM-based conservation score predictors in identifying biologically important motif instances.
• The method demonstrates the ability to quantify motif conservation over broader phylogenetic distances than traditional MSAs.
• Findings suggest that short linear motifs (SLiMs) might be more conserved than previously estimated by MSA-based metrics.

Conclusions:

• PairK offers a more effective approach to quantifying motif conservation in disordered regions, enhancing the prediction of functional protein-protein interactions.
• This method addresses a key challenge in bioinformatics, improving the reliability of SLiM identification.
• The open-source availability of PairK facilitates its adoption in biological research for mapping interaction networks and generating new hypotheses.