ELF: An Extended-Lagrangian Free Energy Calculation Module for Multiple Molecular Dynamics Engines

Area of Science:

• Computational Chemistry
• Molecular Dynamics Simulations
• Free Energy Calculations

Background:

• Extended adaptive biasing force (eABF) is an efficient collective variable (CV)-based importance-sampling algorithm.
• Current eABF implementations are limited to specific molecular dynamics (MD) engines like NAMD and LAMMPS.
• Broader accessibility of eABF and its variants is needed for diverse MD engines.

Purpose of the Study:

• To present a PLUMED-based implementation of eABF and its generalized form (egABF).
• To enable the use of eABF variants with MD engines such as GROMACS, AMBER, CP2K, and openMM.
• To offer a flexible framework for selecting importance-sampling algorithms.

Main Methods:

• Developed an implementation named extended-Lagrangian free energy calculation (ELF) within the PLUMED framework.
• Designed ELF as a stand-alone gradient estimator.
• Ensured compatibility with various CV-based sampling algorithms.

Main Results:

• ELF provides a PLUMED-based implementation for eABF and egABF.
• The implementation is compatible with multiple MD engines beyond NAMD and LAMMPS.
• ELF can serve as a gradient estimator for other advanced sampling methods.

Conclusions:

• The PLUMED-based ELF implementation significantly broadens the applicability of eABF and egABF.
• ELF offers a versatile tool for researchers to choose optimal importance-sampling methods.
• This work enhances the flexibility and accessibility of advanced free energy calculation techniques in molecular dynamics.