Extending the applicability of the ANI deep learning molecular potential to Sulfur and Halogens.

By Christian Devereux, Justin S. Smith, Katey Huddleston, Kipton Barros, Roman I. Zubatyuk, Olexandr Isayev, Adrian E. Roitberg. Journal of Chemical Theory and Computation, 2020

371 citations recorded by the source corpus.

Machine learning (ML) methods have become powerful, predictive tools in a wide range of applications, such as facial recognition and autonomous vehicles. In the sciences, computational chemists and physicists have been using ML for the prediction of physical phenomena, such as atomistic potential energy surfaces and reaction pathways. Transferable ML potentials, such as ANI-1x, have been developed with the goal of accurately simulating organic molecules containing the chemical elements H, C, N, and O. Here we provide an extension of the ANI-1x model. The new model, dubbed ANI-2x, is trained to three additional chemical elements: S, F, and Cl. Additionally, ANI-2x underwent torsional refinement training to better predict molecular torsion profiles. These new features open a wide range of new applications within organic chemistry and drug development. These seven elements (H, C, N, O, F, Cl, S) make up ~90% of drug like molecules. To show that these additions do not sacrifice accuracy, we have tested this model across a range of organic molecules and applications, including the COMP6 benchmark, dihedral rotations, conformer scoring, and non-bonded interactions. ANI-2x is shown to accurately predict molecular energies compared to DFT with a ~106 factor speedup and a negligible slowdown compared to ANI-1x. The resulting model is a valuable tool for drug development that can potentially replace both quantum calculations and classical force fields for myriad applications.

Subfields: Computational Theory and Mathematics, Materials Chemistry, Physical and Theoretical Chemistry

View Paper: https://www.semanticscholar.org/paper/2ac2632a08792cb961aaf96f485f0846309d7ab4

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Theory and Mathematics

In Materials Chemistry

In Physical and Theoretical Chemistry

Citations in the corpus

OpenMM 8: Molecular Dynamics Simulation with Machine Learning Potentials (2023, positive): “They support only a limited set of elements (seven for ANI-2x, four for ANI-1ccx), and do not support charged molecules.”
Towards universal neural network potential for material discovery applicable to arbitrary combination of 45 elements (2021, background): “Architecture Invariance NN type Scalar sph Vector Tensor BPNN [5] E(3) Invariant MLP X ANI-1 [2] E(3) Invariant MLP X ANI-2x [6] E(3) Invariant MLP X Schrodinger-ANI [7] E(3) Invariant MLP X TensorMol-0.”
Learning local equivariant representations for large-scale atomistic dynamics (2022, methodology): “…linear ACE [12], sGDML [10], GAP [6], a classical force-ﬁeld based on the GAFF functional form [52, 53] as well as two ANI parametrizations [47, 48]: ANI-pretrained refers to a version of ANI that was pre-trained on 8.9 million structures and ﬁne-tuned on this data set, while ANI-2x refers to…”
E(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials (2021, methodology): “We ﬁnd that NequIP signiﬁcantly outperforms invariant GNN-IPs (such as SchNet [9] and DimeNet [11]), shallow neural networks (such as ANI [40]), and kernel-based approaches (such as GAP [5], FCHL19/GPR [39, 41] and sGDML [37]).”