The Engine

The Pocket Identificatoin engine provides automated detection and characterization of binding pockets in protein structures, serving as a foundation for structure-based drug discovery. It identifies potential small-molecule binding sites, quantifies their properties, and evaluates their druggability to guide downstream docking and optimization workflows. By combining geometric analysis with physicochemical profiling, the engine helps researchers pinpoint the most promising target cavities for molecular binding and design.

This engine is most often used in tandem with AutoDock VINA within the Revilico platform, streamlining the process from pocket detection to ligand docking and interaction assessment. Users upload protein structures to automatically generate detailed pocket analyses, rank cavities by druggability, and export ready-to-use coordinates for docking simulations.

The Algorithm

Pocket ID leverages an alpha-sphere–based algorithm to detect and characterize cavities on a protein surface.

• Pocket Detection: Scans the 3D protein structure to identify and rank potential binding cavities.

• Pocket Characterization: Quantifies key geometric and physicochemical properties, including pocket volume, solvent accessibility, and hydrophobic density.

• Druggability Scoring: Computes a composite score that reflects how likely a pocket is to accommodate a drug-like ligand.

• Integration with Docking: Outputs coordinates and rankings for direct use in AutoDock VINA campaigns.

Each identified pocket is described by a comprehensive set of descriptors that collectively inform ligand-binding potential and guide rational design decisions.

Algorithm Validation

FPocket is a widely validated open-source method for cavity detection, benchmarked on extensive datasets of crystallographic protein-ligand complexes. Internal tests within Revilico confirm high correlation between FPocket druggability scores and observed docking success rates in AutoDock VINA simulations. Metrics such as pocket score, hydrophobic density, and volume reliably predict ligand compatibility, and performance is consistent across protein classes, from soluble enzymes to membrane-associated receptors.

Scientific Impact

The Pocket Identification engine enables a deeper understanding of target structure and binding site diversity, allowing researchers to:

• Identify druggable pockets across multiple target conformations.

• Quantify pocket properties to assess ligand compatibility and selectivity.

• Prioritize binding sites for fragment-based or virtual screening campaigns.

• Analyze hydrophobic and polar contributions to guide rational compound design.

• Evaluate how mutations or conformational shifts influence pocket accessibility.

By providing objective, quantitative insight into protein cavities, FPocket enhances the precision of target evaluation and optimizes structure-guided discovery.

Business Impact

The Pocket Identification engine helps teams streamline early discovery workflows by:

• Automating the identification of viable drug-binding pockets before docking.

• Reducing manual curation time for structural analysis and pocket evaluation.

• Improving docking accuracy through informed pocket selection and scoring.

• Increasing efficiency in lead identification by coupling pocket analytics with ligand screening.

With Pocket Identification, researchers can move from protein structure to docking-ready models in minutes, improving target prioritization and accelerating the pace of computational drug discovery.