Hammock: a hidden Markov model-based peptide clustering algorithm to identify protein-interaction consensus motifs in...

Area of Science:

• Bioinformatics
• Computational Biology
• Molecular Biology

Background:

• Proteins recognize interaction partners via short linear motifs in disordered regions.
• Experimental techniques use peptides to mimic these motifs, generating large datasets.
• Processing these vast peptide sequence datasets is crucial for studying protein-protein interactions.

Purpose of the Study:

• To develop a software tool for rapid identification of shared specificity motifs in massive peptide datasets.
• To enable large-scale screening and analysis of protein-protein interactions.
• To process and cluster vast amounts of peptide sequences from various experimental sources.

Main Methods:

• Development of the Hammock software tool for sequence clustering and alignment.
• Application of Hammock to datasets of SH3 domain ligands and antibody epitopes.
• Computational feasibility testing for processing extremely large datasets.

Main Results:

• Hammock rapidly identifies clusters of sequences with shared specificity motifs in massive datasets.
• The tool successfully clustered sequences mimicking antibody epitopes, revealing tolerance for variations.
• Computational analysis confirmed the feasibility of processing significantly larger datasets.

Conclusions:

• Hammock is an effective tool for analyzing large-scale peptide sequence data.
• The software facilitates the discovery of protein-protein interaction motifs and antibody epitope variations.
• This approach is computationally scalable for future large-scale interaction studies.