High-performance QM/MM Enhanced Sampling Molecular Dynamics Simulations with GENESIS SPDYN and QSimulate-QM

Area of Science:

• Computational Chemistry
• Molecular Dynamics
• Biophysics

Background:

• Accurate molecular simulations require integrating quantum mechanics (QM) for reactive regions and molecular mechanics (MM) for the environment.
• Existing QM/MM methods face challenges in computational efficiency and handling periodic boundary conditions.

Purpose of the Study:

• To develop and implement a novel QM/MM module within the GENESIS molecular dynamics program.
• To enhance the computational performance and applicability of QM/MM simulations, particularly for enhanced sampling techniques.

Main Methods:

• Integration of QSimulate-QM (electronic structure) with SPDYN (molecular dynamics) in GENESIS.
• Implementation of periodic boundary conditions using real-space QM calculations and Particle Mesh Ewald (PME) for MM charges.
• Optimization of density functional tight-binding (DFTB) calculations and multipole expansions for QM/MM interactions.
• Application of enhanced sampling methods: generalized replica exchange solute tempering (gREST), replica-exchange umbrella sampling (REUS), and path sampling.

Main Results:

• Achieved high performance for QM/MM-MD simulations (over 1 ns/day on a single node) with QM regions up to ~100 atoms and MM regions up to ~100,000 atoms.
• Successfully computed the Ramachandran plot of alanine dipeptide using QM/MM-MD.
• Calculated the potential of mean force (PMF) for a proton transfer reaction in an enzyme at the QM/MM level.

Conclusions:

• The developed QM/MM module significantly enhances the efficiency and capabilities of molecular dynamics simulations.
• This new tool enables advanced sampling methods for studying complex chemical processes in biological systems.
• Facilitates atomistic-level investigations of enzymatic reactions and molecular conformations.