PEP-SiteFinder

PEP-SiteFinder predicts peptide binding sites on protein surfaces by generating de novo three-dimensional peptide conformations and performing blind docking to identify candidate interaction patches.

Key Features:

• 3D De Novo Generation of Peptide Conformations: Generates multiple three-dimensional peptide conformations from sequence for downstream docking.
• Blind Rigid Docking Procedure: Performs fast blind rigid docking of peptide conformations across the entire target protein surface without requiring prior binding-site information.
• Benchmark Validation: Identifies candidate interaction patches overlapping experimentally determined sites in approximately 90% of cases across a benchmark of 41 complexes.
• Propensity Index for Prediction Confidence: Outputs a propensity index that quantifies the confidence of each predicted binding site.
• Coarse-Grained Representation: Employs a coarse-grained representation for proteins and peptides during the search to reduce computational cost while preserving interaction-relevant features.
• Extension to Semi-Flexible Protocols: Can be extended to semi-flexible protocols in which peptide conformations adapt dynamically near potential binding regions.

Scientific Applications:

• Experimental design and validation: Provides candidate binding sites to guide experimental validation of peptide-protein interactions.
• Computational modeling and in silico screening: Serves as a starting point for building computational models and conducting in silico analyses of peptide-protein interfaces.
• Drug development and peptide therapeutics: Supports mapping of peptide-protein interfaces relevant to peptide-based drug discovery and therapeutic development.
• Molecular biology and biochemistry studies: Aids mechanistic and biochemical characterization of peptide-protein binding sites in molecular biology and biochemistry research.

Methodology:

Generates multiple de novo three-dimensional peptide conformations and systematically docks them onto the protein surface using blind rigid docking with optional semi-flexible protocols, employing coarse-grained representations and heuristic methods.