Neural Network Potential with Multi-Resolution Approach Enables Accurate Prediction of Reaction Free Energies in Solution
arXiv:2411.19728 · doi:10.1021/jacs.4c17015
Abstract
We present design and implementation of a novel neural network potential (NNP) and its combination with an electrostatic embedding scheme, commonly used within the context of hybrid quantum-mechanical/molecular-mechanical (QM/MM) simulations. Substitution of a computationally expensive QM Hamiltonian by a NNP with the same accuracy largely reduces the computational cost and enables efficient sampling in prospective MD simulations, the main limitation faced by traditional QM/MM set-ups. The model relies on the recently introduced anisotropic message passing (AMP) formalism to compute atomic interactions and encode symmetries found in QM systems. AMP is shown to be highly efficient in terms of both data and computational costs, and can be readily scaled to sample systems involving more than 350 solute and 40'000 solvent atoms for hundreds of nanoseconds using umbrella sampling. The performance and broad applicability of our approach are showcased by calculating the free-energy surface of alanine dipeptide, the preferred ligation states of nickel phosphine complexes, and dissociation free energies of charged pyridine and quinoline dimers. Results with this ML/MM approach show excellent agreement with experimental data. In contrast, free energies calculated with static high-level QM calculations paired with implicit solvent models or QM/MM MD simulations using cheaper semi-empirical methods show up to ten times higher deviation from the experimental ground truth and sometimes even fail to reproduce qualitative trends.
References in corpus (26)
- scikit-image: Image processing in Python
- SchNet - a deep learning architecture for molecules and materials
- ANI-1: An extensible neural network potential with DFT accuracy at force field computational cost
- E(3)-Equivariant Graph Neural Networks for Data-Efficient and Accurate Interatomic Potentials
- Statistically optimal analysis of samples from multiple equilibrium states
- Correlation consistent valence basis sets for use with the Stuttgart-Dresden-Bonn relativistic effective core potentials: the atoms Ga-Kr and In-Xe
- Machine Learning of Accurate Energy-Conserving Molecular Force Fields
- PhysNet: A Neural Network for Predicting Energies, Forces, Dipole Moments and Partial Charges
- Big Data meets Quantum Chemistry Approximations: The -Machine Learning Approach
- Combining Machine Learning and Computational Chemistry for Predictive Insights Into Chemical Systems
- Machine Learning Unifies the Modelling of Materials and Molecules
- Directional Message Passing for Molecular Graphs
- SpookyNet: Learning Force Fields with Electronic Degrees of Freedom and Nonlocal Effects
- Non-covalent interactions across organic and biological subsets of chemical space: Physics-based potentials parametrized from machine learning
- Forces are not Enough: Benchmark and Critical Evaluation for Machine Learning Force Fields with Molecular Simulations
- Machine Learning in QM/MM Molecular Dynamics Simulations of Condensed-Phase Systems
- The Design Space of E(3)-Equivariant Atom-Centered Interatomic Potentials
- SchNetPack 2.0: A neural network toolbox for atomistic machine learning
- MACE-OFF: Transferable Short Range Machine Learning Force Fields for Organic Molecules
- Universal QM/MM Approaches for General Nanoscale Applications
- TorchMD-Net 2.0: Fast Neural Network Potentials for Molecular Simulations
- Some Observations on the Performance of the Most Recent Exchange-Correlation Functionals for the Large and Chemically Diverse GMTKN55 Benchmark
- PhysNet Meets CHARMM: A Framework for Routine Machine Learning / Molecular Mechanics Simulations
- Learning Local Equivariant Representations for Large-Scale Atomistic Dynamics
- Hybrid Classical/Machine-Learning Force Fields for the Accurate Description of Molecular Condensed-Phase Systems
- Incompleteness of graph neural networks for points clouds in three dimensions