Development of scalable and generalizable machine learned force field for polymers
Abstract Understanding and predicting the properties of polymers is vital to developing tailored polymer molecules for desired applications. Classical force fields may fail to capture key properties, for example, the transport properties of certain polymer systems such as polyethylene glycol. As a s...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2023-10-01
|
| Series: | Scientific Reports |
| Online Access: | https://doi.org/10.1038/s41598-023-43804-5 |
| _version_ | 1797452566988062720 |
|---|---|
| author | Shaswat Mohanty James Stevenson Andrea R. Browning Leif Jacobson Karl Leswing Mathew D. Halls Mohammad Atif Faiz Afzal |
| author_facet | Shaswat Mohanty James Stevenson Andrea R. Browning Leif Jacobson Karl Leswing Mathew D. Halls Mohammad Atif Faiz Afzal |
| author_sort | Shaswat Mohanty |
| collection | DOAJ |
| description | Abstract Understanding and predicting the properties of polymers is vital to developing tailored polymer molecules for desired applications. Classical force fields may fail to capture key properties, for example, the transport properties of certain polymer systems such as polyethylene glycol. As a solution, we present an alternative potential energy surface, a charge recursive neural network (QRNN) model trained on DFT calculations made on smaller atomic clusters that generalizes well to oligomers comprising larger atomic clusters or longer chains. We demonstrate the validity of the polymer QRNN workflow by modeling the oligomers of ethylene glycol. We apply two rounds of active learning (addition of new training clusters based on current model performance) and implement a novel model training approach that uses partial charges from a semi-empirical method. Our developed QRNN model for polymers produces stable molecular dynamics (MD) simulation trajectory and captures the dynamics of polymer chains as indicated by the striking agreement with experimental values. Our model allows working on much larger systems than allowed by DFT simulations, at the same time providing a more accurate force field than classical force fields which provides a promising avenue for large-scale molecular simulations of polymeric systems. |
| first_indexed | 2024-03-09T15:10:32Z |
| format | Article |
| id | doaj.art-2dbe2c1c2a9b4039905d9d1df98e14c1 |
| institution | Directory Open Access Journal |
| issn | 2045-2322 |
| language | English |
| last_indexed | 2024-03-09T15:10:32Z |
| publishDate | 2023-10-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj.art-2dbe2c1c2a9b4039905d9d1df98e14c12023-11-26T13:23:50ZengNature PortfolioScientific Reports2045-23222023-10-0113111510.1038/s41598-023-43804-5Development of scalable and generalizable machine learned force field for polymersShaswat Mohanty0James Stevenson1Andrea R. Browning2Leif Jacobson3Karl Leswing4Mathew D. Halls5Mohammad Atif Faiz Afzal6Schrödinger, Inc.Schrödinger, Inc.Schrödinger, Inc.Schrödinger, Inc.Schrödinger, Inc.Schrödinger, Inc.Schrödinger, Inc.Abstract Understanding and predicting the properties of polymers is vital to developing tailored polymer molecules for desired applications. Classical force fields may fail to capture key properties, for example, the transport properties of certain polymer systems such as polyethylene glycol. As a solution, we present an alternative potential energy surface, a charge recursive neural network (QRNN) model trained on DFT calculations made on smaller atomic clusters that generalizes well to oligomers comprising larger atomic clusters or longer chains. We demonstrate the validity of the polymer QRNN workflow by modeling the oligomers of ethylene glycol. We apply two rounds of active learning (addition of new training clusters based on current model performance) and implement a novel model training approach that uses partial charges from a semi-empirical method. Our developed QRNN model for polymers produces stable molecular dynamics (MD) simulation trajectory and captures the dynamics of polymer chains as indicated by the striking agreement with experimental values. Our model allows working on much larger systems than allowed by DFT simulations, at the same time providing a more accurate force field than classical force fields which provides a promising avenue for large-scale molecular simulations of polymeric systems.https://doi.org/10.1038/s41598-023-43804-5 |
| spellingShingle | Shaswat Mohanty James Stevenson Andrea R. Browning Leif Jacobson Karl Leswing Mathew D. Halls Mohammad Atif Faiz Afzal Development of scalable and generalizable machine learned force field for polymers Scientific Reports |
| title | Development of scalable and generalizable machine learned force field for polymers |
| title_full | Development of scalable and generalizable machine learned force field for polymers |
| title_fullStr | Development of scalable and generalizable machine learned force field for polymers |
| title_full_unstemmed | Development of scalable and generalizable machine learned force field for polymers |
| title_short | Development of scalable and generalizable machine learned force field for polymers |
| title_sort | development of scalable and generalizable machine learned force field for polymers |
| url | https://doi.org/10.1038/s41598-023-43804-5 |
| work_keys_str_mv | AT shaswatmohanty developmentofscalableandgeneralizablemachinelearnedforcefieldforpolymers AT jamesstevenson developmentofscalableandgeneralizablemachinelearnedforcefieldforpolymers AT andrearbrowning developmentofscalableandgeneralizablemachinelearnedforcefieldforpolymers AT leifjacobson developmentofscalableandgeneralizablemachinelearnedforcefieldforpolymers AT karlleswing developmentofscalableandgeneralizablemachinelearnedforcefieldforpolymers AT mathewdhalls developmentofscalableandgeneralizablemachinelearnedforcefieldforpolymers AT mohammadatiffaizafzal developmentofscalableandgeneralizablemachinelearnedforcefieldforpolymers |