Divide-and-link peptide docking: a fragment-based peptide docking protocol.

Area of Science:

• Computational biology
• Structural biology
• Biochemistry

Background:

• Protein-peptide interactions are vital for cellular processes, protein folding, and peptide drug development.
• Experimental structural data is limited, necessitating computational prediction of binding modes.

Purpose of the Study:

• To design and evaluate a novel fragment-based computational protocol, Divide-and-Link Peptide Docking (DLPepDock), for predicting protein-peptide binding modes.

Main Methods:

• DLPepDock fragments peptides, docks them individually, links poses, filters conformations, and scores using MM/PBSA.
• Utilizes fragmental coordinate transformation, geometrical filters, clustering, and molecular force field minimization.
• Employs Amber for binding energy calculations and third-party tools for fragment docking.

Main Results:

• DLPepDock demonstrated superior performance compared to 15 other docking protocols on the LEADS-PEP benchmark dataset.
• The protocol accurately predicted binding modes for protein-peptide complexes with peptides of 6-12 residues.
• Outperformed established methods like FlexPepDock ab initio and AutoDock CrankPep in accuracy.

Conclusions:

• DLPepDock offers a robust and accurate computational approach for predicting protein-peptide interaction modes.
• This method can aid in understanding fundamental biological processes and accelerate peptide drug design.
• The protocol's effectiveness is validated against a diverse benchmark dataset, highlighting its potential utility.