Machine Learning of Coarse-Grained Molecular Dynamics Force Fields

  title={Machine Learning of Coarse-Grained Molecular Dynamics Force Fields},
  author={Jiang Wang and Christoph Wehmeyer and Frank No{\'e} and Cecilia Clementi},
  journal={ACS Central Science},
  pages={755 - 767}
Atomistic or ab initio molecular dynamics simulations are widely used to predict thermodynamics and kinetics and relate them to molecular structure. A common approach to go beyond the time- and length-scales accessible with such computationally expensive simulations is the definition of coarse-grained molecular models. Existing coarse-graining approaches define an effective interaction potential to match defined properties of high-resolution models or experimental data. In this paper, we… 

Figures from this paper

Ensemble learning of coarse-grained molecular dynamics force fields with a kernel approach.

This work proposes a 2-layer training scheme that enables GDML to learn an effective coarse-grained (CG) model from all-atom simulation data in a sample efficient manner and yields a smaller free energy error than neural networks when the training set is small, and a comparably high accuracy when theTraining set is sufficiently large.

Coarse-graining auto-encoders for molecular dynamics

This work proposes a generative modeling framework based on variational auto-encoders to unify the tasks of learning discrete coarse-grained variables, decoding back to atomistic detail, and parameterizing coarse- grained force fields.

Statistically Optimal Force Aggregation for Coarse-Graining Molecular Dynamics.

An optimization statement for force mappings is defined and it is demonstrated that substantially improved CG force fields can be learned from the same simulation data when using optimized force maps.

Developing Machine-Learned Potentials for Coarse-Grained Molecular Simulations: Challenges and Pitfalls

This work explores the application of SchNet models to obtain a CG potential for liquid benzene, investigating the effect of model architecture and hyperparameters on the thermodynamic, dynamical, and structural properties of the simulated CG systems, reporting and discussing challenges encountered and future directions envisioned.

Simulate Time-integrated Coarse-grained Molecular Dynamics with Geometric Machine Learning

A novel score-based GNN refinement module resolves the long-standing challenge of long-time simulation instability and can generalize to unseen novel systems and simulate for much longer than the training trajectories.

Coarse graining molecular dynamics with graph neural networks.

A hybrid architecture for the machine learning of coarse-grained force fields that learn their own features via a subnetwork that leverages continuous filter convolutions on a graph neural network architecture is introduced that succeeds at reproducing the thermodynamics for small biomolecular systems.

Machine Learning Coarse-Grained Potentials of Protein Thermodynamics

It is shown that a single coarse-grained potential can integrate all twelve proteins and can capture experimental structural features of mutated proteins, indicating that machine learning coarse- grained potentials could provide a feasible approach to simulate and understand protein dynamics.

Anisotropic molecular coarse-graining by force and torque matching with neural networks

We develop a machine-learning method for coarse-graining condensed-phase molecular systems using anisotropic particles. The method extends currently available high-dimensional neural network

Ensuring thermodynamic consistency with invertible coarse-graining.

This work uses a graph machine learning framework to embed atomic configurations into a low-dimensional space to produce efficient representations of the original molecular system and shows that this technique is robust, recovering the first two moments of the distribution of several observables in proteins such as chignolin and alanine dipeptide.

Accurate Molecular Dynamics Enabled by Efficient Physically Constrained Machine Learning Approaches

The symmetric GDML (sGDML) approach is able to faithfully reproduce global force fields at the accuracy high-level ab initio methods, thus enabling sample intensive tasks like molecular dynamics simulations at that level of accuracy.

Neural Network Based Prediction of Conformational Free Energies - A New Route toward Coarse-Grained Simulation Models.

A new method where a neural network is used to extract high-dimensional free energy surfaces (FES) from molecular dynamics simulation trajectories and it is shown that the NN not only is able to correctly describe the free-energy surface for oligomer lengths that it was trained on but also was able to predict the conformational sampling of longer chains.

Towards exact molecular dynamics simulations with machine-learned force fields

A flexible machine-learning force-field with high-level accuracy for molecular dynamics simulations is developed, for flexible molecules with up to a few dozen atoms and insights into the dynamical behavior of these molecules are provided.

Learning Effective Molecular Models from Experimental Observables.

It is shown that when the correct coarsening resolution is used not only do the optimized models match the Reference model simulated experimental data accurately but additional observables not directly targeted during the optimization procedure are also reproduced.

Coarse-graining entropy, forces, and structures.

The present work investigates parallels between these seemingly divergent approaches by examining the relative entropy and multiscale coarse-graining (MS-CG) methods, and demonstrates that the MS-CG method minimizes the average of its gradient squared.

Machine-Learned Coarse-Grained Models.

This work presents a novel FF optimization framework that integrates MD simulations with particle swarm optimization (PSO) algorithm and artificial neural network (ANN), which was used to develop transferable coarse-grained models for D2O and DMF as a proof of concept.

Machine learning of accurate energy-conserving molecular force fields

The GDML approach enables quantitative molecular dynamics simulations for molecules at a fraction of cost of explicit AIMD calculations, thereby allowing the construction of efficient force fields with the accuracy and transferability of high-level ab initio methods.

Many-Body Coarse-Grained Interactions Using Gaussian Approximation Potentials.

We introduce a computational framework that is able to describe general many-body coarse-grained (CG) interactions of molecules and use it to model the free energy surface of molecular liquids as a

DeePCG: Constructing coarse-grained models via deep neural networks.

It is found that the two-body, three- body, and higher-order oxygen correlation functions produced by the coarse-grained and full atomistic models agree very well with each other, illustrating the effectiveness of the DeePCG model on a rather challenging task.

The multiscale coarse-graining method. I. A rigorous bridge between atomistic and coarse-grained models.

The present work develops a formal statistical mechanical framework for the MS-CG method and demonstrates that the variational principle underlying the method may, in principle, be employed to determine the many-body potential of mean force (PMF) that governs the equilibrium distribution of positions of the CG sites for theMS-CG models.

Coarse-graining methods for computational biology.

The theoretical underpinnings and history of coarse-grained methods are discussed and the state of the field is summarized, organizing key methodologies based on an emerging paradigm for multiscale theory and modeling of biomolecular systems.