AlloPred

AlloPred predicts allosteric pockets on protein structures to identify sites involved in functional regulation and allosteric drug targeting.

Key Features:

• Normal Mode Analysis: Perturbs normal modes using normal mode analysis to identify regions susceptible to allosteric regulation.
• Pocket Descriptors: Computes pocket descriptors that characterize structural features of protein cavities for ranking potential allosteric sites.
• Machine Learning Integration: Combines normal mode perturbations and pocket descriptors with machine learning algorithms to rank pockets by likelihood of functional relevance.
• Performance Evaluation: In benchmarks on 40 known allosteric proteins, it ranked an allosteric pocket as the top candidate in 23 cases and included an allosteric pocket in the top two ranks in 28 of 40 cases.

Scientific Applications:

• Drug Discovery: Identifying candidate allosteric sites for experimental validation and allosteric drug development.
• Understanding Allostery: Mapping potential allosteric sites to inform studies of protein signaling and regulation.

Methodology:

AlloPred perturbs normal modes via normal mode analysis to identify susceptible regions, computes pocket descriptors for candidate cavities, and applies machine learning to rank potential allosteric pockets.