Science Brief: Binding Free Energy Redefined: Accurate, Fast, Scalable

Binding Free Energy Redefined: Accurate, Fast, and Scalable

Chris Neale, Gaetano Calabró, Agnes Huang, and Lukas Eberlein

OpenEye, Cadence Molecular Sciences, 9 Bisbee Court, Santa Fe, NM 87508

Summary:

OpenEye's Free Energy Nonequilibrium Switching (FE-NES) accurately ranks ligands by their relative binding free energies to accelerate lead optimization.
Calculate hundreds to thousands of compounds in a few hours with cloud-scale parallelization.
Maximize throughput and speed with efficient simulations.

Product Keywords: Orion^®, MD, Free Energy Nonequilibrium Switching, FE-NES

Abstract:

Computational estimation of ligand-protein binding free energies is essential in the lead optimization phase of drug discovery for both industry and academia. Industry scientists have reported that OpenEye’s Free Energy Non-Equilibrium Switching (FE-NES) offers accurate ranking while delivering better speed and scale compared to traditional methods like free energy perturbation and thermodynamic integration, which typically require extensive sampling near equilibrium.

FE-NES implementation on the Orion^® molecular design cloud platform provides a ready-to-use solution. This enables automated calculations, from hundreds to thousands of compounds, in just a few hours. It is an ideal tool for daily lead optimization tasks.

To evaluate free energy prediction accuracy, we use the Kendall’s tau rank-order correlation coefficient, which measures how closely predicted results align with experimental data. FE-NES delivers accuracy on par with industry-leading solutions.

FE-NES delivers best in class speed and best in class scale while rank-ordering molecules to provide decision support that is competitive with any other physics-based affinity prediction tool. Data sets originally curated in [1, 2].

Users can adjust FE-NES switches to balance accuracy and throughput, optimizing hundreds to thousands of molecules in real time to meet tight drug discovery project timelines. With various map types (OELOMAP, multi-star map, etc.) and an interactive interface, users can leverage their expertise to make confident design decisions without the typical computational bottleneck.

[1] Wang et al., 2015, J. Am. Chem. Soc., 137: 2695.

[2] Schindler et al., 2020, J. Chem. Inf. Model., 60: 5457