Exploring the Potential of Parallel Biasing in Flat Histogram Methods

Abstract

Metadynamics, a member of the `flat histogram' class of advanced sampling algorithms, has been widely used in molecular simulations to drive the exploration of states separated by high free energy barriers and promote comprehensive sampling of free energy landscapes defined on collective variables (CVs) which characterize the state of the system. Typically, the methods encounter severe limitations when exploring large numbers of CVs. A recently proposed variant, parallel bias metadynamics (PBMetaD), promises to aid in exploring free energy landscapes along with multiple important collective variables by exchanging the $n$-dimensional free energy landscape required by standard methods for $n$ one-dimensional marginal free energy landscapes. In this study, we systematically examine how parallel biasing affects the convergence of free energy landscapes along with each variable relative to standard methods and the effectiveness of the parallel biasing strategy for addressing common bottlenecks in the use of advanced sampling to calculate free energies.

arXiv commentSH and JKW acknowledge support for this project from the United States National Science Foundation (Award No. DMR-1751988). MJQ and ECM were supported by the MICCoM Center at Argonne National Laboratory under a Computational Materials Science center grant from the Department of Energy, Basic Energy Sciences Division

References in corpus (9)

• Escaping free-energy minima
• PLUMED 2: New feathers for an old bird
• Well-Tempered Metadynamics: A Smoothly Converging and Tunable Free-Energy Method
• An efficient, multiple range random walk algorithm to calculate the density of states
• String Method for the Study of Rare Events
• Determining the density of states for classical statistical models: A random walk algorithm to produce a flat histogram
• Time-lagged autoencoders: Deep learning of slow collective variables for molecular kinetics
• Molecular enhanced sampling with autoencoders: On-the-fly collective variable discovery and accelerated free energy landscape exploration
• Nonlinear Discovery of Slow Molecular Modes using State-Free Reversible VAMPnets