Machine-Learned Molecular Surface and Its Application to Implicit Solvent Simulations

Wei, Haixin; Zhao, Zekai; Luo, Ray

doi:10.1021/acs.jctc.1c00492

Cited by 4 publications

(10 citation statements)

References 83 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Several efficient finite-difference numerical solvers, both linear − and nonlinear, are implemented in pbsa for various applications of the Poisson–Boltzmann method. The GPU support of those solvers is also implemented in pbsa.cuda . − In the 2023 release, improvements to the pbsa program include the integration of the Machine-Learned Solvent Excluded Surface (MLSES) model, which provides a highly efficient and differentiable molecular surface for continuum solvation modeling of biomolecules. Various options for the MLSES model have been implemented, allowing users to optimize performance on both central processing unit (CPU) and GPU platforms using Fortran, the CUDA kernel, and LibTorch.…”

Section: Introductionmentioning

confidence: 99%

AmberTools

Case,

Aktulga,

Belfon

et al. 2023

J. Chem. Inf. Model.

Self Cite

301

121

View full text Add to dashboard Cite

show abstract

Section: Introductionmentioning

confidence: 99%

AmberTools

Case,

Aktulga,

Belfon

et al. 2023

J. Chem. Inf. Model.

Self Cite

301

121

View full text Add to dashboard Cite

show abstract

“…It is then not surprising that the recent explosive development of machine learning (ML) techniques, including deep neural networks (DNNs), − is already making a noticeable impact on this field, − including works aimed directly at improving the accuracy of the description of complex solvation effects. − It should be noted that the majority of these recent works combine quantum mechanics (QM)-based methodology with ML, while our interest here is purely classical approaches. Among recent purely ML-based approaches, a featurization algorithm, based on functional class fingerprints and implemented within the DeepChem ML framework, was used in ref to predict the hydration-free energies (HFEs) of a diverse set of 642 neutral small molecules available in FreeSolvarguably the largest public database of experimentally measured HFEs.…”

Section: Introductionmentioning

confidence: 99%

Improving the Accuracy of Physics-Based Hydration-Free Energy Predictions by Machine Learning the Remaining Error Relative to the Experiment

Bass,

Elder,

Folescu

et al. 2023

J. Chem. Theory Comput.

View full text Add to dashboard Cite

The accuracy of computational models of water is key to atomistic simulations of biomolecules. We propose a computationally efficient way to improve the accuracy of the prediction of hydration-free energies (HFEs) of small molecules: the remaining errors of the physics-based models relative to the experiment are predicted and mitigated by machine learning (ML) as a postprocessing step. Specifically, the trained graph convolutional neural network attempts to identify the "blind spots" in the physics-based model predictions, where the complex physics of aqueous solvation is poorly accounted for, and partially corrects for them. The strategy is explored for five classical solvent models representing various accuracy/speed trade-offs, from the fast analytical generalized Born (GB) to the popular TIP3P explicit solvent model; experimental HFEs of small neutral molecules from the FreeSolv set are used for the training and testing. For all of the models, the ML correction reduces the resulting root-mean-square error relative to the experiment for HFEs of small molecules, without significant overfitting and with negligible computational overhead. For example, on the test set, the relative accuracy improvement is 47% for the fast analytical GB, making it, after the ML correction, almost as accurate as uncorrected TIP3P. For the TIP3P model, the accuracy improvement is about 39%, bringing the ML-corrected model's accuracy below the 1 kcal/mol threshold. In general, the relative benefit of the ML corrections is smaller for more accurate physics-based models, reaching the lower limit of about 20% relative accuracy gain compared with that of the physics-based treatment alone. The proposed strategy of using ML to learn the remaining error of physics-based models offers a distinct advantage over training ML alone directly on reference HFEs: it preserves the correct overall trend, even well outside of the training set.

show abstract

“…Additionally, results are sensitive to grid discretization. , It is worth noting that a surface-free Poisson–Boltzmann solver model treats the solute and solvent uniformly, bypassing the necessity of generating a molecular surface . The level set function, a mathematical tool leveraged in computer graphics, has displayed versatility in shape representation and analysis. , Although the efficiency has improved compared with analytical algorithms, it is still far from ready for deployment in drug screening in terms of speed. Lately, the application of machine learning techniques has garnered increasing attention due to their flexibility and efficiency in fitting given a sufficient number of data samples.…”

mentioning

confidence: 99%

“…Lately, the application of machine learning techniques has garnered increasing attention due to their flexibility and efficiency in fitting given a sufficient number of data samples. Successes have been documented in various disciplines, including chemistry and physics. − Considering the rapidly improving computational performance of hardware (e.g., TPU (tensor processing unit) and GPU (graphics processing unit)), the enhanced utilization of these techniques is projected to boost efficiency.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Grid-Robust Efficient Neural Interface Model for Universal Molecule Surface Construction from Point Clouds

Wu,

Wei,

Zhu

et al. 2023

J. Phys. Chem. Lett.

Self Cite

View full text Add to dashboard Cite

Molecular surfaces play a pivotal role in elucidating the properties and functions of biological complexes. While various surfaces have been proposed for specific scenarios, their widespread adoption faces challenges due to limited efficiency stemming from hand-crafted modeling designs. In this work, we proposed a general framework that incorporates both the point cloud concept and neural networks. The use of matrix multiplication in this framework enables efficient implementation across diverse platforms and libraries. We applied this framework to develop the GENIUSES (Grid-robust Efficient Neural Interface for Universal Solvent-Excluded Surface) model for constructing SES. GENIUSES demonstrates high accuracy and efficiency across data sets with varying conformations and complexities. Compared to the classical implementation of SES in the AMBER software package, our framework achieved a 26-fold speedup while retaining ∼95% accuracy when ported to the GPU platform using CUDA. Greater speedups can be obtained in large-scale systems. Importantly, our model exhibits robustness against variations in the grid spacing. We have integrated this infrastructure into AMBER to enhance accessibility for research in drug screening and related fields, where efficiency is of paramount importance.

show abstract

Machine-Learned Molecular Surface and Its Application to Implicit Solvent Simulations

Cited by 4 publications

References 83 publications

AmberTools

AmberTools

Improving the Accuracy of Physics-Based Hydration-Free Energy Predictions by Machine Learning the Remaining Error Relative to the Experiment

Grid-Robust Efficient Neural Interface Model for Universal Molecule Surface Construction from Point Clouds

Contact Info

Product

Resources

About