Incorporating Neural Networks into the AMOEBA Polarizable Force Field

Overview

Journal J Phys Chem B

Publisher American Chemical Society

Specialty Chemistry

Date 2024 Mar 6

PMID 38445577

Authors

Yanxing Wang

Theo Jaffrelot Inizan

Chengwen Liu

Jean-Philip Piquemal

Pengyu Ren

Affiliations

Soon will be listed here.

Abstract

Neural network potentials (NNPs) offer significant promise to bridge the gap between the accuracy of quantum mechanics and the efficiency of molecular mechanics in molecular simulation. Most NNPs rely on the locality assumption that ensures the model's transferability and scalability and thus lack the treatment of long-range interactions, which are essential for molecular systems in the condensed phase. Here we present an integrated hybrid model, AMOEBA+NN, which combines the AMOEBA potential for the short- and long-range noncovalent atomic interactions and an NNP to capture the remaining local covalent contributions. The AMOEBA+NN model was trained on the conformational energy of the ANI-1x data set and tested on several external data sets ranging from small molecules to tetrapeptides. The hybrid model demonstrated substantial improvements over the baseline models in term of accuracy as the molecule size increased, suggesting its potential as a next-generation approach for chemically accurate molecular simulations.

References

Zubatyuk R, Smith J, Leszczynski J, Isayev O . Accurate and transferable multitask prediction of chemical properties with an atoms-in-molecules neural network. Sci Adv. 2019; 5(8):eaav6490. PMC: 6688864. DOI: 10.1126/sciadv.aav6490. View

Inizan T, Ple T, Adjoua O, Ren P, Gokcan H, Isayev O . Scalable hybrid deep neural networks/polarizable potentials biomolecular simulations including long-range effects. Chem Sci. 2023; 14(20):5438-5452. PMC: 10208042. DOI: 10.1039/d2sc04815a. View

Ple T, Lagardere L, Piquemal J . Force-field-enhanced neural network interactions: from local equivariant embedding to atom-in-molecule properties and long-range effects. Chem Sci. 2023; 14(44):12554-12569. PMC: 10646944. DOI: 10.1039/d3sc02581k. View

Najibi A, Goerigk L . The Nonlocal Kernel in van der Waals Density Functionals as an Additive Correction: An Extensive Analysis with Special Emphasis on the B97M-V and ωB97M-V Approaches. J Chem Theory Comput. 2018; 14(11):5725-5738. DOI: 10.1021/acs.jctc.8b00842. View

Behler J, Parrinello M . Generalized neural-network representation of high-dimensional potential-energy surfaces. Phys Rev Lett. 2007; 98(14):146401. DOI: 10.1103/PhysRevLett.98.146401. View

Christensen A, Sirumalla S, Qiao Z, OConnor M, Smith D, Ding F . OrbNet Denali: A machine learning potential for biological and organic chemistry with semi-empirical cost and DFT accuracy. J Chem Phys. 2021; 155(20):204103. DOI: 10.1063/5.0061990. View

Anstine D, Isayev O . Machine Learning Interatomic Potentials and Long-Range Physics. J Phys Chem A. 2023; 127(11):2417-2431. PMC: 10041642. DOI: 10.1021/acs.jpca.2c06778. View

Walker B, Liu C, Wait E, Ren P . Automation of AMOEBA polarizable force field for small molecules: Poltype 2. J Comput Chem. 2022; 43(23):1530-1542. PMC: 9329217. DOI: 10.1002/jcc.26954. View

Zgarbova M, Rosnik A, Luque F, Curutchet C, Jurecka P . Transferability and additivity of dihedral parameters in polarizable and nonpolarizable empirical force fields. J Comput Chem. 2015; 36(25):1874-84. DOI: 10.1002/jcc.24012. View

10.

Wu J, Chattree G, Ren P . Automation of AMOEBA polarizable force field parameterization for small molecules. Theor Chem Acc. 2012; 131(3):1138. PMC: 3322661. DOI: 10.1007/s00214-012-1138-6. View

11.

Wang H, Yang W . Force Field for Water Based on Neural Network. J Phys Chem Lett. 2018; 9(12):3232-3240. PMC: 6241215. DOI: 10.1021/acs.jpclett.8b01131. View

12.

Mardirossian N, Head-Gordon M . ωB97M-V: A combinatorially optimized, range-separated hybrid, meta-GGA density functional with VV10 nonlocal correlation. J Chem Phys. 2016; 144(21):214110. DOI: 10.1063/1.4952647. View

13.

Unke O, Chmiela S, Sauceda H, Gastegger M, Poltavsky I, Schutt K . Machine Learning Force Fields. Chem Rev. 2021; 121(16):10142-10186. PMC: 8391964. DOI: 10.1021/acs.chemrev.0c01111. View

14.

Smith J, Isayev O, Roitberg A . ANI-1: an extensible neural network potential with DFT accuracy at force field computational cost. Chem Sci. 2017; 8(4):3192-3203. PMC: 5414547. DOI: 10.1039/c6sc05720a. View

15.

Wang Y, Walker B, Liu C, Ren P . An Efficient Approach to Large-Scale Ab Initio Conformational Energy Profiles of Small Molecules. Molecules. 2022; 27(23). PMC: 9738817. DOI: 10.3390/molecules27238567. View

16.

Unke O, Meuwly M . PhysNet: A Neural Network for Predicting Energies, Forces, Dipole Moments, and Partial Charges. J Chem Theory Comput. 2019; 15(6):3678-3693. DOI: 10.1021/acs.jctc.9b00181. View

17.

Rackers J, Wang Z, Lu C, Laury M, Lagardere L, Schnieders M . Tinker 8: Software Tools for Molecular Design. J Chem Theory Comput. 2018; 14(10):5273-5289. PMC: 6335969. DOI: 10.1021/acs.jctc.8b00529. View

18.

Lubbers N, Smith J, Barros K . Hierarchical modeling of molecular energies using a deep neural network. J Chem Phys. 2018; 148(24):241715. DOI: 10.1063/1.5011181. View

19.

Ren P, Wu C, Ponder J . Polarizable Atomic Multipole-based Molecular Mechanics for Organic Molecules. J Chem Theory Comput. 2011; 7(10):3143-3161. PMC: 3196664. DOI: 10.1021/ct200304d. View

20.

Lagardere L, Jolly L, Lipparini F, Aviat F, Stamm B, Jing Z . Tinker-HP: a massively parallel molecular dynamics package for multiscale simulations of large complex systems with advanced point dipole polarizable force fields. Chem Sci. 2018; 9(4):956-972. PMC: 5909332. DOI: 10.1039/c7sc04531j. View