用机器学习代理模型代替分子动力学计算加速多尺度力场参数优化。

IF 2.2 3区 化学 Q3 CHEMISTRY, PHYSICAL
Robin Strickstrock, Alexander Hagg, Dirk Reith, Karl N Kirschner
{"title":"用机器学习代理模型代替分子动力学计算加速多尺度力场参数优化。","authors":"Robin Strickstrock, Alexander Hagg, Dirk Reith, Karl N Kirschner","doi":"10.1002/cphc.202500353","DOIUrl":null,"url":null,"abstract":"<p><p>Molecular modeling plays a vital role in many scientific fields, ranging from material science to drug design. To predict and investigate the properties of those systems, a suitable force field (FF) is required. Improving the accuracy or expanding the applicability of the FFs is an ongoing process, referred to as force-field parameter (FFParam) optimization. In recent years, data-driven machine learning (ML) algorithms have become increasingly relevant in computational sciences and elevated the capability of many molecular modeling methods. Herein, time-consuming molecular dynamic simulations, used during a multiscale FFParam optimization, are substituted by a ML surrogate model to speed-up the optimization process. Subject to this multiscale optimization are the Lennard-Jones parameters for carbon and hydrogen that are used to reproduce the target properties: n-octane's relative conformational energies and its bulk-phase density. By substituting the most time-consuming element of this optimization, the required time is reduced by a factor of ≈20, while retaining FFs with similar quality. Furthermore, the workflow used to obtain the surrogate model (i.e., training data acquisition, data preparation, model selection, and training) for such substitution is presented.</p>","PeriodicalId":9819,"journal":{"name":"Chemphyschem","volume":" ","pages":"e2500353"},"PeriodicalIF":2.2000,"publicationDate":"2025-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Speed up Multi-Scale Force-Field Parameter Optimization by Substituting Molecular Dynamics Calculations with a Machine Learning Surrogate Model.\",\"authors\":\"Robin Strickstrock, Alexander Hagg, Dirk Reith, Karl N Kirschner\",\"doi\":\"10.1002/cphc.202500353\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><p>Molecular modeling plays a vital role in many scientific fields, ranging from material science to drug design. To predict and investigate the properties of those systems, a suitable force field (FF) is required. Improving the accuracy or expanding the applicability of the FFs is an ongoing process, referred to as force-field parameter (FFParam) optimization. In recent years, data-driven machine learning (ML) algorithms have become increasingly relevant in computational sciences and elevated the capability of many molecular modeling methods. Herein, time-consuming molecular dynamic simulations, used during a multiscale FFParam optimization, are substituted by a ML surrogate model to speed-up the optimization process. Subject to this multiscale optimization are the Lennard-Jones parameters for carbon and hydrogen that are used to reproduce the target properties: n-octane's relative conformational energies and its bulk-phase density. By substituting the most time-consuming element of this optimization, the required time is reduced by a factor of ≈20, while retaining FFs with similar quality. Furthermore, the workflow used to obtain the surrogate model (i.e., training data acquisition, data preparation, model selection, and training) for such substitution is presented.</p>\",\"PeriodicalId\":9819,\"journal\":{\"name\":\"Chemphyschem\",\"volume\":\" \",\"pages\":\"e2500353\"},\"PeriodicalIF\":2.2000,\"publicationDate\":\"2025-09-05\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Chemphyschem\",\"FirstCategoryId\":\"92\",\"ListUrlMain\":\"https://doi.org/10.1002/cphc.202500353\",\"RegionNum\":3,\"RegionCategory\":\"化学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q3\",\"JCRName\":\"CHEMISTRY, PHYSICAL\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Chemphyschem","FirstCategoryId":"92","ListUrlMain":"https://doi.org/10.1002/cphc.202500353","RegionNum":3,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"CHEMISTRY, PHYSICAL","Score":null,"Total":0}
引用次数: 0

摘要

分子建模在许多科学领域起着至关重要的作用,从材料科学到药物设计。为了预测和研究这些系统的性质,需要一个合适的力场(FF)。提高FFs的精度或扩大其适用性是一个持续的过程,称为力场参数(FFParam)优化。近年来,数据驱动的机器学习(ML)算法在计算科学中越来越重要,并提高了许多分子建模方法的能力。在此,在多尺度FFParam优化中使用的耗时的分子动力学模拟被ML代理模型取代,以加快优化过程。受这种多尺度优化影响的是碳和氢的Lennard-Jones参数,这些参数用于再现目标性质:正辛烷的相对构象能及其体相密度。通过替换该优化中最耗时的元素,所需时间减少了约20倍,同时保留了类似质量的FFs。此外,还介绍了用于获取替代模型的工作流(即,训练数据获取、数据准备、模型选择和训练)。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
Speed up Multi-Scale Force-Field Parameter Optimization by Substituting Molecular Dynamics Calculations with a Machine Learning Surrogate Model.

Molecular modeling plays a vital role in many scientific fields, ranging from material science to drug design. To predict and investigate the properties of those systems, a suitable force field (FF) is required. Improving the accuracy or expanding the applicability of the FFs is an ongoing process, referred to as force-field parameter (FFParam) optimization. In recent years, data-driven machine learning (ML) algorithms have become increasingly relevant in computational sciences and elevated the capability of many molecular modeling methods. Herein, time-consuming molecular dynamic simulations, used during a multiscale FFParam optimization, are substituted by a ML surrogate model to speed-up the optimization process. Subject to this multiscale optimization are the Lennard-Jones parameters for carbon and hydrogen that are used to reproduce the target properties: n-octane's relative conformational energies and its bulk-phase density. By substituting the most time-consuming element of this optimization, the required time is reduced by a factor of ≈20, while retaining FFs with similar quality. Furthermore, the workflow used to obtain the surrogate model (i.e., training data acquisition, data preparation, model selection, and training) for such substitution is presented.

求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
Chemphyschem
Chemphyschem 化学-物理:原子、分子和化学物理
CiteScore
4.60
自引率
3.40%
发文量
425
审稿时长
1.1 months
期刊介绍: ChemPhysChem is one of the leading chemistry/physics interdisciplinary journals (ISI Impact Factor 2018: 3.077) for physical chemistry and chemical physics. It is published on behalf of Chemistry Europe, an association of 16 European chemical societies. ChemPhysChem is an international source for important primary and critical secondary information across the whole field of physical chemistry and chemical physics. It integrates this wide and flourishing field ranging from Solid State and Soft-Matter Research, Electro- and Photochemistry, Femtochemistry and Nanotechnology, Complex Systems, Single-Molecule Research, Clusters and Colloids, Catalysis and Surface Science, Biophysics and Physical Biochemistry, Atmospheric and Environmental Chemistry, and many more topics. ChemPhysChem is peer-reviewed.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:604180095
Book学术官方微信