BsFinder

BsFinder predicts binding sites on protein 3D structures to support molecular docking, de novo drug design, and the structural identification and comparison of functional sites.

Key Features:

• Heuristic approach: Employs a novel heuristic method to identify similar binding sites through direct comparison of protein 3D structures.
• Local sequence alignment: Uses local sequence alignment to identify regions with potential functional similarity.
• Protein surface detection: Detects and analyzes protein surface features where binding sites are likely located.
• 3D structure comparison: Compares 3D structures of protein surfaces to find similar binding sites across different proteins.
• Implementation: Implemented in Java.
• Evaluation on PDB: Evaluated by comparing a complete protein against all complete protein structures in the Protein Data Bank (PDB).
• Reported performance: Reports an average recall of 82% on PDB comparisons, compared with typical recall values below 50% for other approaches.

Scientific Applications:

• Molecular docking: Identifies binding sites to inform docking studies and ligand placement.
• De novo drug design: Provides binding-site information relevant for structure-based drug design efforts.
• Structural identification of functional sites: Supports identification and comparison of functional sites across proteins.
• Annotation of unknown structures: Assists in annotating protein structures with unknown functions by detecting putative binding sites.

Methodology:

Implemented in Java and applying a novel heuristic approach that comprises local sequence alignment, protein surface detection, and comparison of protein 3D structures.