Learning molecular potentials with neural networks

Abstract

The potential energy of molecular species and their conformers can be computed with a wide range of computational chemistry methods, from molecular mechanics to ab initio quantum chemistry. However, the proper choice of the computational approach based on computational cost and reliability of calculated energies is a dilemma, especially for large molecules. This dilemma is proved to be even more problematic for studies that require hundreds and thousands of calculations, such as drug discovery. On the other hand, driven by their pattern recognition capabilities, neural networks started to gain popularity in the computational chemistry community. During the last decade, many neural network potentials have been developed to predict a variety of chemical information of different systems. Neural network potentials are proved to predict chemical properties with accuracy comparable to quantum mechanical approaches but with the cost approaching molecular mechanics calculations. As a result, the development of more reliable, transferable, and extensible neural network potentials became an attractive field of study for researchers. In this review, we outlined an overview of the status of current neural network potentials and strategies to improve their accuracy. We provide recent examples of studies that prove the applicability of these potentials. We also discuss the capabilities and shortcomings of the current models and the challenges and future aspects of their development and applications. It is expected that this review would provide guidance for the development of neural network potentials and the exploitation of their applicability. This article is categorized under: Data Science > Artificial Intelligence/Machine Learning Molecular and Statistical Mechanics > Molecular Interactions Software > Molecular Modeling

Prediction of protein pKawith representation learning

“…These potentials, such as ANI 91–98 and AIMNet, 99 can learn the electronic environment of an atom in conjunction with the many-body symmetry functions that arise from the coordinates. 100,101 Using this learned information and combining it with the structural fingerprints that depend on the coordinates, NNPs can predict target molecular properties such as energy and forces. Thus, NNPs can be utilized to obtain information that stems from the atomic environment, and this information can be used to train ML models for protein p K a estimations.…”

Molecular Property Prediction and Molecular Design Using a Supervised Grammar Variational Autoencoder

“…Here, we summarized the most important features of the selected molecular databases to test the SGVAE performance, while additional computational details can be found elsewhere, e.g., QM9, ,, PubChemQC, , and QM7-X. , …”

Prediction of Protein pKa with Representation Learning

“…Minimized structures are used as inputs for NNP to compute all descriptors. A detailed description of ANI neural network potential and corresponding descriptors can be found in elsewhere 96,100 .…”

“…Over the last decade, NNPs have been shown to provide accuracy approaching of QM calculations and comparable computational cost with all-atom force fields. These potentials, such as ANI [91][92][93][94][95][96][97][98] and AIMNet 99 , can learn the electronic environment of an atom in conjunction with the many-body symmetry functions that arise from the coordinates 100,101 . Using this learned information and combining it with the structural fingerprints that depend on the coordinates, NNPs can predict target molecular properties such as energy and forces.…”