AutoDock CrankPep: combining folding and docking to predict protein-peptide complexes

Area of Science:

• Computational Biology
• Structural Biology
• Drug Discovery

Background:

• Protein-peptide interactions are crucial for cellular functions.
• Predicting these interactions is vital for therapeutic development.
• Existing docking software struggles with peptide size and flexibility.

Purpose of the Study:

• To introduce AutoDock CrankPep (ADCP), a novel computational approach for docking flexible peptides into rigid protein receptors.
• To evaluate ADCP's performance against established peptide docking methods.

Main Methods:

• ADCP models protein-peptide complexes by folding peptides within the protein's potential field.
• The method was tested on standard benchmarking datasets (LEADS-PEP, peptiDB) and longer peptides (16-20 amino acids).

Main Results:

• ADCP outperforms leading peptide docking methods on benchmark datasets.
• The approach demonstrates reliable docking for peptides up to 20 amino acids.
• ADCP enables accurate modeling of peptide-protein and peptide-disordered protein interactions.

Conclusions:

• ADCP offers a robust solution for predicting protein-peptide complexes, overcoming limitations of existing software.
• This method facilitates the accurate modeling of complex biological interactions.
• ADCP is an open-source tool, promoting accessibility for researchers.