Cargando…

End-to-end differentiable construction of molecular mechanics force fields

Molecular mechanics (MM) potentials have long been a workhorse of computational chemistry. Leveraging accuracy and speed, these functional forms find use in a wide variety of applications in biomolecular modeling and drug discovery, from rapid virtual screening to detailed free energy calculations....

Descripción completa

Detalles Bibliográficos
Autores principales: Wang, Yuanqing, Fass, Josh, Kaminow, Benjamin, Herr, John E., Rufa, Dominic, Zhang, Ivy, Pulido, Iván, Henry, Mike, Bruce Macdonald, Hannah E., Takaba, Kenichiro, Chodera, John D.
Formato: Online Artículo Texto
Lenguaje:English
Publicado: The Royal Society of Chemistry 2022
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9600499/
https://www.ncbi.nlm.nih.gov/pubmed/36349096
http://dx.doi.org/10.1039/d2sc02739a
_version_ 1784816857839042560
author Wang, Yuanqing
Fass, Josh
Kaminow, Benjamin
Herr, John E.
Rufa, Dominic
Zhang, Ivy
Pulido, Iván
Henry, Mike
Bruce Macdonald, Hannah E.
Takaba, Kenichiro
Chodera, John D.
author_facet Wang, Yuanqing
Fass, Josh
Kaminow, Benjamin
Herr, John E.
Rufa, Dominic
Zhang, Ivy
Pulido, Iván
Henry, Mike
Bruce Macdonald, Hannah E.
Takaba, Kenichiro
Chodera, John D.
author_sort Wang, Yuanqing
collection PubMed
description Molecular mechanics (MM) potentials have long been a workhorse of computational chemistry. Leveraging accuracy and speed, these functional forms find use in a wide variety of applications in biomolecular modeling and drug discovery, from rapid virtual screening to detailed free energy calculations. Traditionally, MM potentials have relied on human-curated, inflexible, and poorly extensible discrete chemical perception rules (atom types) for applying parameters to small molecules or biopolymers, making it difficult to optimize both types and parameters to fit quantum chemical or physical property data. Here, we propose an alternative approach that uses graph neural networks to perceive chemical environments, producing continuous atom embeddings from which valence and nonbonded parameters can be predicted using invariance-preserving layers. Since all stages are built from smooth neural functions, the entire process—spanning chemical perception to parameter assignment—is modular and end-to-end differentiable with respect to model parameters, allowing new force fields to be easily constructed, extended, and applied to arbitrary molecules. We show that this approach is not only sufficiently expressive to reproduce legacy atom types, but that it can learn to accurately reproduce and extend existing molecular mechanics force fields. Trained with arbitrary loss functions, it can construct entirely new force fields self-consistently applicable to both biopolymers and small molecules directly from quantum chemical calculations, with superior fidelity than traditional atom or parameter typing schemes. When adapted to simultaneously fit partial charge models, espaloma delivers high-quality partial atomic charges orders of magnitude faster than current best-practices with low inaccuracy. When trained on the same quantum chemical small molecule dataset used to parameterize the Open Force Field (“Parsley”) openff-1.2.0 small molecule force field augmented with a peptide dataset, the resulting espaloma model shows superior accuracy vis-á-vis experiments in computing relative alchemical free energy calculations for a popular benchmark. This approach is implemented in the free and open source package espaloma, available at https://github.com/choderalab/espaloma.
format Online
Article
Text
id pubmed-9600499
institution National Center for Biotechnology Information
language English
publishDate 2022
publisher The Royal Society of Chemistry
record_format MEDLINE/PubMed
spelling pubmed-96004992022-11-07 End-to-end differentiable construction of molecular mechanics force fields Wang, Yuanqing Fass, Josh Kaminow, Benjamin Herr, John E. Rufa, Dominic Zhang, Ivy Pulido, Iván Henry, Mike Bruce Macdonald, Hannah E. Takaba, Kenichiro Chodera, John D. Chem Sci Chemistry Molecular mechanics (MM) potentials have long been a workhorse of computational chemistry. Leveraging accuracy and speed, these functional forms find use in a wide variety of applications in biomolecular modeling and drug discovery, from rapid virtual screening to detailed free energy calculations. Traditionally, MM potentials have relied on human-curated, inflexible, and poorly extensible discrete chemical perception rules (atom types) for applying parameters to small molecules or biopolymers, making it difficult to optimize both types and parameters to fit quantum chemical or physical property data. Here, we propose an alternative approach that uses graph neural networks to perceive chemical environments, producing continuous atom embeddings from which valence and nonbonded parameters can be predicted using invariance-preserving layers. Since all stages are built from smooth neural functions, the entire process—spanning chemical perception to parameter assignment—is modular and end-to-end differentiable with respect to model parameters, allowing new force fields to be easily constructed, extended, and applied to arbitrary molecules. We show that this approach is not only sufficiently expressive to reproduce legacy atom types, but that it can learn to accurately reproduce and extend existing molecular mechanics force fields. Trained with arbitrary loss functions, it can construct entirely new force fields self-consistently applicable to both biopolymers and small molecules directly from quantum chemical calculations, with superior fidelity than traditional atom or parameter typing schemes. When adapted to simultaneously fit partial charge models, espaloma delivers high-quality partial atomic charges orders of magnitude faster than current best-practices with low inaccuracy. When trained on the same quantum chemical small molecule dataset used to parameterize the Open Force Field (“Parsley”) openff-1.2.0 small molecule force field augmented with a peptide dataset, the resulting espaloma model shows superior accuracy vis-á-vis experiments in computing relative alchemical free energy calculations for a popular benchmark. This approach is implemented in the free and open source package espaloma, available at https://github.com/choderalab/espaloma. The Royal Society of Chemistry 2022-09-08 /pmc/articles/PMC9600499/ /pubmed/36349096 http://dx.doi.org/10.1039/d2sc02739a Text en This journal is © The Royal Society of Chemistry https://creativecommons.org/licenses/by/3.0/
spellingShingle Chemistry
Wang, Yuanqing
Fass, Josh
Kaminow, Benjamin
Herr, John E.
Rufa, Dominic
Zhang, Ivy
Pulido, Iván
Henry, Mike
Bruce Macdonald, Hannah E.
Takaba, Kenichiro
Chodera, John D.
End-to-end differentiable construction of molecular mechanics force fields
title End-to-end differentiable construction of molecular mechanics force fields
title_full End-to-end differentiable construction of molecular mechanics force fields
title_fullStr End-to-end differentiable construction of molecular mechanics force fields
title_full_unstemmed End-to-end differentiable construction of molecular mechanics force fields
title_short End-to-end differentiable construction of molecular mechanics force fields
title_sort end-to-end differentiable construction of molecular mechanics force fields
topic Chemistry
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9600499/
https://www.ncbi.nlm.nih.gov/pubmed/36349096
http://dx.doi.org/10.1039/d2sc02739a
work_keys_str_mv AT wangyuanqing endtoenddifferentiableconstructionofmolecularmechanicsforcefields
AT fassjosh endtoenddifferentiableconstructionofmolecularmechanicsforcefields
AT kaminowbenjamin endtoenddifferentiableconstructionofmolecularmechanicsforcefields
AT herrjohne endtoenddifferentiableconstructionofmolecularmechanicsforcefields
AT rufadominic endtoenddifferentiableconstructionofmolecularmechanicsforcefields
AT zhangivy endtoenddifferentiableconstructionofmolecularmechanicsforcefields
AT pulidoivan endtoenddifferentiableconstructionofmolecularmechanicsforcefields
AT henrymike endtoenddifferentiableconstructionofmolecularmechanicsforcefields
AT brucemacdonaldhannahe endtoenddifferentiableconstructionofmolecularmechanicsforcefields
AT takabakenichiro endtoenddifferentiableconstructionofmolecularmechanicsforcefields
AT choderajohnd endtoenddifferentiableconstructionofmolecularmechanicsforcefields