QuantumBind-RBFE: Accurate Relative Binding Free Energy Calculations Using Neural Network Potentials DOI
Francesc Sabanés Zariquiey, Stephen E. Farr, Stefan H. Doerr

и другие.

Journal of Chemical Information and Modeling, Год журнала: 2025, Номер unknown

Опубликована: Апрель 8, 2025

Accurate prediction of protein-ligand binding affinities is crucial in drug discovery, particularly during hit-to-lead and lead optimization phases, however, limitations ligand force fields continue to impact accuracy. In this work, we validate relative free energy (RBFE) accuracy using neural network potentials (NNPs) for the ligands. We utilize a novel NNP model, AceFF 1.0, based on TensorNet architecture small molecules that broadens applicability diverse drug-like compounds, including all important chemical elements supporting charged molecules. Using established benchmarks, show overall improved correlation affinity predictions compared with GAFF2 molecular mechanics ANI2-x NNPs. Slightly less but comparable correlations OPLS4. also can run simulations at 2 fs time step, least two times larger than previous models, providing significant speed gains. The results promise further evolutions calculations NNPs while demonstrating its practical use already current generation. code model are publicly available research use.

Язык: Английский

QuantumBind-RBFE: Accurate Relative Binding Free Energy Calculations Using Neural Network Potentials DOI
Francesc Sabanés Zariquiey, Stephen E. Farr, Stefan H. Doerr

и другие.

Journal of Chemical Information and Modeling, Год журнала: 2025, Номер unknown

Опубликована: Апрель 8, 2025

Accurate prediction of protein-ligand binding affinities is crucial in drug discovery, particularly during hit-to-lead and lead optimization phases, however, limitations ligand force fields continue to impact accuracy. In this work, we validate relative free energy (RBFE) accuracy using neural network potentials (NNPs) for the ligands. We utilize a novel NNP model, AceFF 1.0, based on TensorNet architecture small molecules that broadens applicability diverse drug-like compounds, including all important chemical elements supporting charged molecules. Using established benchmarks, show overall improved correlation affinity predictions compared with GAFF2 molecular mechanics ANI2-x NNPs. Slightly less but comparable correlations OPLS4. also can run simulations at 2 fs time step, least two times larger than previous models, providing significant speed gains. The results promise further evolutions calculations NNPs while demonstrating its practical use already current generation. code model are publicly available research use.

Язык: Английский

Процитировано

0