EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Deep learning path-like collective variable for enhanced sampling molecular dynamics

Thorben Fröhlking iD, Luigi Bonati iD, Valerio Rizzi iD, Francesco Luigi Gervasio iD

The Journal of Chemical Physics · 2024

DOI: 10.1063/5.0202156 ↗

Kurz gefasst

Fachartikel in der Zeitschrift The Journal of Chemical Physics, 2024. Verfasst von Thorben Fröhlking und 3 weiteren Autor:innen. Beteiligte Einrichtung: School of Pharmaceutical Sciences, University of Geneva, Rue Michel Servet 1, 1206 Genève, Switzerland. Hauptthemen laut MEDLINE: Molecular Dynamics Simulation, Deep Learning. Die Arbeit wurde bisher 29-mal zitiert (Crossref) und nennt 72 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung, Bundesministerium für Bildung und Forschung.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Several enhanced sampling techniques rely on the definition of collective variables to effectively explore free energy landscapes. The existing variables that describe the progression along a reactive pathway offer an elegant solution but face a number of limitations. In this paper, we address these challenges by introducing a new path-like collective variable called the “deep-locally non-linear-embedding,” which is inspired by principles of the locally linear embedding technique and is trained on a reactive trajectory. The variable mimics the ideal reaction coordinate by automatically generating a non-linear combination of features through a differentiable generalized autoencoder that combines a neural network with a continuous k-nearest neighbor selection. Among the key advantages of this method is its capability to automatically choose the metric for searching neighbors and to learn the path from state A to state B without the need to handpick landmarks a priori. We demonstrate the effectiveness of DeepLNE by showing that the progression along the path variable closely approximates the ideal reaction coordinate in toy models, such as the Müller-Brown potential and alanine dipeptide. Then, we use it in the molecular dynamics simulations of an RNA tetraloop, where we highlight its capability to accelerate transitions and estimate the free energy of folding.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Molecular Dynamics SimulationDeep LearningRNAThermodynamicsDipeptides

Substanzen

RNADipeptides

Autor:innen und Institutionen

  • School of Pharmaceutical Sciences, University of Geneva, Rue Michel Servet 1, 1206 Genève, Switzerland.
  • Institute of Pharmaceutical Sciences of Western Switzerland (ISPSO), University of Geneva, 1206 Genève, Switzerland.
  • Swiss Institute of Bioinformatics, University of Geneva, 1206 Genève, Switzerland.
  • Italian Institute of Technology, Via Melen 83, 16152 Genoa, Italy.
  • Department of Chemistry, University College London, London WC1E 6BT, United Kingdom.

ORCID-Profile: Thorben Fröhlking, Luigi Bonati, Valerio Rizzi, Francesco Luigi Gervasio

Förderung

  • Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung
  • Bundesministerium für Bildung und Forschung

Zitieren

APA

Fröhlking, T., Bonati, L., Rizzi, V., & Gervasio, F. L. (2024). Deep learning path-like collective variable for enhanced sampling molecular dynamics. The Journal of Chemical Physics. https://doi.org/10.1063/5.0202156

Vancouver

Fröhlking T, Bonati L, Rizzi V, Gervasio FL. Deep learning path-like collective variable for enhanced sampling molecular dynamics. The Journal of Chemical Physics. 2024. doi:10.1063/5.0202156

BibTeX
@article{Frhlking2024,
  author = {Fröhlking, Thorben and Bonati, Luigi and Rizzi, Valerio and Gervasio, Francesco Luigi},
  title = {Deep learning path-like collective variable for enhanced sampling molecular dynamics},
  journal = {The Journal of Chemical Physics},
  year = {2024},
  doi = {10.1063/5.0202156}
}
Kostenlos starten →