On Using Atomistic Solvent Layers in Hybrid All-Atom/Coarse-Grained Molecular Dynamics Simulations.

Area of Science:

• Computational Chemistry
• Biomolecular Simulations
• Molecular Dynamics

Background:

• Hybrid all-atom/coarse-grained (AA-CG) simulations offer computational speed-up for large biomolecular systems.
• Coarse-graining solvent reduces computational cost, but atomic solvent layers are crucial for local solvation effects.

Purpose of the Study:

• To devise and validate fixed-resolution AA-CG schemes with and without atomic water layers.
• To investigate the impact of CG solvent type and AA solvent region size on simulation accuracy.

Main Methods:

• Calculated Gibbs free energies of solvation for amino acid side chain analogues.
• Systematically validated AA-CG schemes, analyzing effects of distance restraints and CG solvent properties (polarizable vs. nonpolarizable).

Main Results:

• Distance restraints for AA solvent can cause artificial density increases and affect solvation free energies.
• Shifting distance restraints away from the solute significantly improves accuracy.
• Achieved mean unsigned errors of 2.3 kJ/mol (polarizable CG) and 2.6 kJ/mol (nonpolarizable CG) with respect to experimental data.

Conclusions:

• Optimizing the AA-CG boundary is critical for accurate hybrid simulations.
• The nonpolarizable CG water model offers a computationally attractive speed-up.
• This work provides solutions for modeling resolution boundaries in hybrid AA-CG simulations.