Machine learning force fields and coarse-grained variables in molecular dynamics: application to materials and biological systems

Gkeka, Paraskevi; Stoltz, Gabriel; Farimani, Amir Barati; Belkacemi, Zineb; Ceriotti, Michele; Chodera, John D.; Dinner, Aaron R.; Ferguson, Andrew L.; Maillet, Jean‐Bernard; Minoux, Hervé; Peter, Christine; Pietrucci, Fabio; Silveira, Ana Cristina Martins; Tkatchenko, Alexandre; Trstanova, Zofia; Wiewiora, Rafal; Lelièvre, Tony

doi:10.48550/arxiv.2004.06950

Cited by 6 publications

(6 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our method is computationally inexpensive and simulation times are short. The parameter searching process using MBAR analysis can be further automated and optimized using machine learning algorithms, 127,128 thus making our method applicable for exhaustive comparison studies that examine the effect of varying external conditions (ligand concentration, initial conformations, etc.). A reliable backmapping algorithm 129−131 combined with a careful temperature-scaling scheme 132 can be used for converting the Cαbased trajectories into the atomic resolution ensembles, opening the way for a powerful multi-scale exploration of biological systems.…”

Section: Resultsmentioning

confidence: 99%

Coarse-Grained Modeling of Multiple Pathways in Conformational Transitions of Multi-Domain Proteins

Shinobu

Kobayashi

Matsunaga

et al. 2021

J. Chem. Inf. Model.

View full text Add to dashboard Cite

Large-scale conformational transitions in multi-domain proteins are often essential for their functions. To investigate the transitions, it is necessary to explore multiple potential pathways, which involve different intermediate structures. Here, we present a multi-basin (MB) coarse-grained (CG) structure-based Go̅ model for describing transitions in proteins with more than two moving domains. This model is an extension of our dual-basin Go̅ model in which system-dependent parameters are determined systematically using the multistate Bennett acceptance ratio method. In the MB Go̅ model for multi-domain proteins, we assume that intermediate structures may have partial inter-domain native contacts. This approach allows us to search multiple transition pathways that involve distinct intermediate structures using the CG molecular dynamics (MD) simulations. We apply this scheme to an enzyme, adenylate kinase (AdK), which has three major domains and can move along two different pathways. Using the optimized mixing parameters for each pathway, AdK shows frequent transitions between the Open, Closed, and the intermediate basins and samples a wide variety of conformations within each basin. The explored multiple transition pathways could be compared with experimental data and examined in more detail by atomistic MD simulations.

show abstract

Section: Resultsmentioning

confidence: 99%

Coarse-Grained Modeling of Multiple Pathways in Conformational Transitions of Multi-Domain Proteins

Shinobu

Kobayashi

Matsunaga

et al. 2021

J. Chem. Inf. Model.

View full text Add to dashboard Cite

show abstract

“…The lessons learned to build ML models in chemistry and materials science largely transfer to soft matter and biomolecules, where similar constraints on the representation prevail [186]. Screening studies that make use of kernel-based ML have become prominent, for instance in protein-ligand binding, but many typically use experimental data [176].…”

Section: Discussionmentioning

confidence: 99%

Computational compound screening of biomolecules and soft materials by molecular simulations

Bereau¹

2020

Preprint

View full text Add to dashboard Cite

Decades of hardware, methodological, and algorithmic development have propelled molecular dynamics (MD) simulations to the forefront of materials-modeling techniques, bridging the gap between electronic-structure theory and continuum methods. The physics-based approach makes MD appropriate to study emergent phenomena, but simultaneously incurs significant computational investment. This topical review explores the use of MD outside the scope of individual systems, but rather considering many compounds. Such an in silico screening approach makes MD amenable to establishing coveted structure-property relationships. We specifically focus on biomolecules and soft materials, characterized by the significant role of entropic contributions and heterogeneous systems and scales. An account of the state of the art for the implementation of an MD-based screening paradigm is described, including automated force-field parametrization, system preparation, and efficient sampling across both conformation and composition. Emphasis is placed on machinelearning methods to enable MD-based screening. The resulting framework enables the generation of compound-property databases and the use of advanced statistical modeling to gather insight. The review further summarizes a number of relevant applications.

show abstract

“…In other words, 𝜉(𝑞) ∈ ℳ is the macroscopic state of a microscopic state 𝑞 ∈ 𝒟. Designing a good reaction coordinate is a difficult problem, that will not be discussed further in the present work (see [12] for a recent review on the question of automatic learning of transition coordinates). The free energy associated to 𝜉 is then expressed as follows: for every 𝑧 ∈ ℳ,…”

Section: Metastability Reaction Coordinate and Free-energy Profilesmentioning

confidence: 99%

The adaptive biasing force algorithm with non-conservative forces and related topics

2022

View full text Add to dashboard Cite

We propose a study of the Adaptive Biasing Force method's robustness under generic (possibly non-conservative) forces. We first ensure the flat histogram property is satisfied in all cases. We then introduce a fixed point problem yielding the existence of a stationary state for both the Adaptive Biasing Force and Projected Adapted Biasing Force algorithms, relying on generic bounds on the invariant probability measures of homogeneous diffusions. Using classical entropy techniques, we prove the exponential convergence of both biasing force and law as time goes to infinity, for both the Adaptive Biasing Force and the Projected Adaptive Biasing Force methods.

show abstract

Machine learning force fields and coarse-grained variables in molecular dynamics: application to materials and biological systems

Cited by 6 publications

References 0 publications

Coarse-Grained Modeling of Multiple Pathways in Conformational Transitions of Multi-Domain Proteins

Coarse-Grained Modeling of Multiple Pathways in Conformational Transitions of Multi-Domain Proteins

Computational compound screening of biomolecules and soft materials by molecular simulations

The adaptive biasing force algorithm with non-conservative forces and related topics

Contact Info

Product

Resources

About