The Engine

The Flexible and Ensemble Docking engine is a GPU-accelerated virtual screening platform built for rapid prediction of small molecule–protein interactions. It enables researchers to screen thousands of compounds in parallel, reducing computation times from hours to minutes while maintaining precision and reproducibility. Designed for flexibility, the engine supports multiple docking modes and integrates seamlessly with Revilico’s Data Engineering module for compound and receptor management.

Users can upload compound libraries in SMILES format and protein structures in PDB format, configure binding site parameters, and execute high-performance docking campaigns directly through the Command Center. The system automatically handles data validation, progress tracking, and output management for streamlined hit discovery.

The Algorithm

The engine offers three docking methodologies tailored for different levels of structural complexity:

• Rigid Receptor Docking: Treats the protein structure as fixed, ideal for fast initial screening across large compound libraries.

• Flexible Docking: Allows user-defined residues to move during docking, providing more accurate predictions of binding interactions in partially dynamic regions.

• Ensemble Docking: Incorporates receptor conformations derived from molecular dynamics simulations, capturing protein flexibility and enhancing prediction accuracy for challenging targets.

Docking parameters such as box coordinates, dimensions, and exhaustiveness can be precisely defined, while residue validation ensures torsional limits are respected. Parallel GPU processing distributes docking tasks efficiently across available resources, delivering large-scale screening results at exceptional speed.

Algorithm Validation

The Flexible and Ensemble Docking engine has been benchmarked against standard docking suites including AutoDock Vina and Glide, demonstrating comparable or superior binding energy predictions with significantly reduced runtime. Rigid docking results reproduce known ligand-binding orientations within 2 Å RMSD, while flexible and ensemble modes correlate strongly with experimentally observed binding affinities. Internal testing across diverse protein classes confirms robust accuracy and scalability, even for large datasets exceeding 10,000 compounds.

Scientific Impact

This engine provides an essential foundation for structure-based drug discovery:

• Enables rapid identification of high-affinity compounds through large-scale screening.

• Supports flexible and ensemble docking to capture protein dynamics.

• Integrates seamlessly with molecular dynamics and target analytics workflows for end-to-end hit evaluation.

• Provides high-resolution output data suitable for post-screening analysis, binding pose clustering, and structure-activity modeling.

By allowing researchers to combine speed, accuracy, and structural insight, the Static Docking Screening engine accelerates the transition from virtual libraries to viable lead compounds.

Business Impact

The Flexible and Ensemble Docking engine helps organizations optimize discovery pipelines by:

• Reducing computational costs through GPU-parallelized efficiency and credit-based usage.

• Accelerating hit identification from weeks to days with automated batch screening.

• Improving decision quality via reproducible, quantitative binding analyses.

• Enabling scalability for libraries ranging from hundreds to tens of thousands of molecules.

Through integrated data management and automation, this engine streamlines early hit discovery, empowering R&D teams to evaluate chemical space efficiently and focus resources on the most promising candidates.