Articles producció científica> Química Física i Inorgànica

Neural network learns physical rules for copolymer translocation through amphiphilic barriers

  • Identification data

    Identifier: imarina:6406085
    Authors:
    Werner MGuo YBaulin VA
    Abstract:
    © 2020, The Author(s). Recent developments in computer processing power lead to new paradigms of how problems in many-body physics and especially polymer physics can be addressed. Parallel processors can be exploited to generate millions of molecular configurations in complex environments at a second, and concomitant free-energy landscapes can be estimated. Databases that are complete in terms of polymer sequences and architecture form a powerful training basis for cross-checking and verifying machine learning-based models. We employ an exhaustive enumeration of polymer sequence space to benchmark the prediction made by a neural network. In our example, we consider the translocation time of a copolymer through a lipid membrane as a function of its sequence of hydrophilic and hydrophobic units. First, we demonstrate that massively parallel Rosenbluth sampling for all possible sequences of a polymer allows for meaningful dynamic interpretation in terms of the mean first escape times through the membrane. Second, we train a multi-layer neural network on logarithmic translocation times and show by the reduction of the training set to a narrow window of translocation times that the neural network develops an internal representation of the physical rules for sequence-controlled diffusion barriers. Based on the narrow training set, the network result approximates the order of magnitude of translocation times in a window that is several orders of magnitude wider than the training window. We investigate how prediction accuracy depends on the distance of unexplored sequences from the training window.
  • Others:

    Author, as appears in the article.: Werner M; Guo Y; Baulin VA
    Department: Química Física i Inorgànica
    URV's Author/s: Baulin, Vladimir
    Keywords: Topology Simulations Prediction Polymer translocation Peptides Nonelectrolyte partition-coefficients Molecules Membrane Macromolecules Glass-transition temperatures
    Abstract: © 2020, The Author(s). Recent developments in computer processing power lead to new paradigms of how problems in many-body physics and especially polymer physics can be addressed. Parallel processors can be exploited to generate millions of molecular configurations in complex environments at a second, and concomitant free-energy landscapes can be estimated. Databases that are complete in terms of polymer sequences and architecture form a powerful training basis for cross-checking and verifying machine learning-based models. We employ an exhaustive enumeration of polymer sequence space to benchmark the prediction made by a neural network. In our example, we consider the translocation time of a copolymer through a lipid membrane as a function of its sequence of hydrophilic and hydrophobic units. First, we demonstrate that massively parallel Rosenbluth sampling for all possible sequences of a polymer allows for meaningful dynamic interpretation in terms of the mean first escape times through the membrane. Second, we train a multi-layer neural network on logarithmic translocation times and show by the reduction of the training set to a narrow window of translocation times that the neural network develops an internal representation of the physical rules for sequence-controlled diffusion barriers. Based on the narrow training set, the network result approximates the order of magnitude of translocation times in a window that is several orders of magnitude wider than the training window. We investigate how prediction accuracy depends on the distance of unexplored sequences from the training window.
    Thematic Areas: Modeling and simulation Mechanics of materials Materials science, multidisciplinary Materials science (miscellaneous) Materials science (all) General materials science Computer science applications Chemistry, physical
    licence for use: https://creativecommons.org/licenses/by/3.0/es/
    ISSN: 20573960
    Author's mail: vladimir.baulin@urv.cat
    Author identifier: 0000-0003-2086-4271
    Record's date: 2023-02-19
    Papper version: info:eu-repo/semantics/publishedVersion
    Link to the original source: https://www.nature.com/articles/s41524-020-0318-5
    Papper original source: Npj Computational Materials. 6 (1):
    APA: Werner M; Guo Y; Baulin VA (2020). Neural network learns physical rules for copolymer translocation through amphiphilic barriers. Npj Computational Materials, 6(1), -. DOI: 10.1038/s41524-020-0318-5
    Licence document URL: https://repositori.urv.cat/ca/proteccio-de-dades/
    Article's DOI: 10.1038/s41524-020-0318-5
    Entity: Universitat Rovira i Virgili
    Journal publication year: 2020
    Publication Type: Journal Publications
  • Keywords:

    Chemistry, Physical,Computer Science Applications,Materials Science (Miscellaneous),Materials Science, Multidisciplinary,Mechanics of Materials,Modeling and Simulation
    Topology
    Simulations
    Prediction
    Polymer translocation
    Peptides
    Nonelectrolyte partition-coefficients
    Molecules
    Membrane
    Macromolecules
    Glass-transition temperatures
    Modeling and simulation
    Mechanics of materials
    Materials science, multidisciplinary
    Materials science (miscellaneous)
    Materials science (all)
    General materials science
    Computer science applications
    Chemistry, physical
  • Documents:

  • Cerca a google

    Search to google scholar