Combining Force Fields and Neural Networks for an Accurate Representation of Chemically Diverse Molecular Interactions.

Autor:	Illarionov A; InterX Inc. (a Subsidiary of NeoTX Therapeutics Ltd.), 805 Allston Way, Berkeley, California 94710, United States., Sakipov S; InterX Inc. (a Subsidiary of NeoTX Therapeutics Ltd.), 805 Allston Way, Berkeley, California 94710, United States., Pereyaslavets L; InterX Inc. (a Subsidiary of NeoTX Therapeutics Ltd.), 805 Allston Way, Berkeley, California 94710, United States., Kurnikov IV; InterX Inc. (a Subsidiary of NeoTX Therapeutics Ltd.), 805 Allston Way, Berkeley, California 94710, United States., Kamath G; InterX Inc. (a Subsidiary of NeoTX Therapeutics Ltd.), 805 Allston Way, Berkeley, California 94710, United States., Butin O; InterX Inc. (a Subsidiary of NeoTX Therapeutics Ltd.), 805 Allston Way, Berkeley, California 94710, United States., Voronina E; InterX Inc. (a Subsidiary of NeoTX Therapeutics Ltd.), 805 Allston Way, Berkeley, California 94710, United States.; Lomonosov MSU, Skobeltsyn Institute of Nuclear Physics, Moscow, 119991, Russia., Ivahnenko I; InterX Inc. (a Subsidiary of NeoTX Therapeutics Ltd.), 805 Allston Way, Berkeley, California 94710, United States., Leontyev I; InterX Inc. (a Subsidiary of NeoTX Therapeutics Ltd.), 805 Allston Way, Berkeley, California 94710, United States., Nawrocki G; InterX Inc. (a Subsidiary of NeoTX Therapeutics Ltd.), 805 Allston Way, Berkeley, California 94710, United States., Darkhovskiy M; InterX Inc. (a Subsidiary of NeoTX Therapeutics Ltd.), 805 Allston Way, Berkeley, California 94710, United States., Olevanov M; InterX Inc. (a Subsidiary of NeoTX Therapeutics Ltd.), 805 Allston Way, Berkeley, California 94710, United States.; Lomonosov MSU, Dept. of Physics, Moscow, 119991, Russia., Cherniavskyi YK; InterX Inc. (a Subsidiary of NeoTX Therapeutics Ltd.), 805 Allston Way, Berkeley, California 94710, United States., Lock C; InterX Inc. (a Subsidiary of NeoTX Therapeutics Ltd.), 805 Allston Way, Berkeley, California 94710, United States.; Department of Neurology and Neurological Sciences, Stanford University School of Medicine, Palo Alto, California 94304, United States., Greenslade S; InterX Inc. (a Subsidiary of NeoTX Therapeutics Ltd.), 805 Allston Way, Berkeley, California 94710, United States., Sankaranarayanan SK; Center for Nanoscale Materials, Argonne National Lab, Argonne, Illinois 604391, United States.; Department of Mechanical and Industrial Engineering, University of Illinois, Chicago, Illinois 60607, United States., Kurnikova MG; Department of Chemistry, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213, United States., Potoff J; Department of Chemical Engineering and Materials Science, Wayne State University, Detroit, Michigan 48202, United States., Kornberg RD; Department of Structural Biology, Stanford University School of Medicine, Stanford, California 94304, United States., Levitt M; Department of Structural Biology, Stanford University School of Medicine, Stanford, California 94304, United States., Fain B; InterX Inc. (a Subsidiary of NeoTX Therapeutics Ltd.), 805 Allston Way, Berkeley, California 94710, United States.
Jazyk:	angličtina
Zdroj:	Journal of the American Chemical Society [J Am Chem Soc] 2023 Nov 01; Vol. 145 (43), pp. 23620-23629. Date of Electronic Publication: 2023 Oct 19.
DOI:	10.1021/jacs.3c07628
Abstrakt:	A key goal of molecular modeling is the accurate reproduction of the true quantum mechanical potential energy of arbitrary molecular ensembles with a tractable classical approximation. The challenges are that analytical expressions found in general purpose force fields struggle to faithfully represent the intermolecular quantum potential energy surface at close distances and in strong interaction regimes; that the more accurate neural network approximations do not capture crucial physics concepts, e.g., nonadditive inductive contributions and application of electric fields; and that the ultra-accurate narrowly targeted models have difficulty generalizing to the entire chemical space. We therefore designed a hybrid wide-coverage intermolecular interaction model consisting of an analytically polarizable force field combined with a short-range neural network correction for the total intermolecular interaction energy. Here, we describe the methodology and apply the model to accurately determine the properties of water, the free energy of solvation of neutral and charged molecules, and the binding free energy of ligands to proteins. The correction is subtyped for distinct chemical species to match the underlying force field, to segment and reduce the amount of quantum training data, and to increase accuracy and computational speed. For the systems considered, the hybrid ab initio parametrized Hamiltonian reproduces the two-body dimer quantum mechanics (QM) energies to within 0.03 kcal/mol and the nonadditive many-molecule contributions to within 2%. Simulations of molecular systems using this interaction model run at speeds of several nanoseconds per day.
Databáze:	MEDLINE
Externí odkaz:	Zobrazit plný text záznamu