Sigmoid Accelerated Molecular Dynamics: An Efficient Enhanced Sampling Method for Biosystems.

Area of Science:

• Computational chemistry
• Biophysics
• Molecular dynamics simulations

Background:

• Enhanced sampling methods like Gaussian accelerated molecular dynamics (GaMD) are crucial for overcoming timescale limitations in biomolecular simulations.
• Balancing high acceleration with accurate reweighting remains a key challenge in these methods.

Purpose of the Study:

• To introduce Sigmoid accelerated molecular dynamics (SaMD), a novel enhanced sampling technique inspired by GaMD.
• To evaluate SaMD's performance in improving the balance between acceleration and reweighting accuracy in molecular dynamics.

Main Methods:

• Development of SaMD by incorporating a Sigmoid boost potential into molecular dynamics simulations.
• Application and comparison of SaMD against GaMD across three distinct simulation tasks: alanine dipeptide, chignolin folding, and protein-ligand binding.

Main Results:

• SaMD demonstrated improved accuracy and efficiency in generating the free energy landscape for alanine dipeptide.
• SaMD achieved faster convergence (approximately 30%) to experimental Gibbs free energy values for chignolin folding.
• SaMD reproduced protein-ligand binding conformations closer to crystal structures and significantly reduced simulation time (approximately 60%) for HIV protease-XK263 binding.

Conclusions:

• SaMD offers a superior balance between acceleration and reweighting accuracy compared to GaMD.
• SaMD effectively extends the accessible timescales and enhances computational efficiency in biomolecular simulations.
• The proposed SaMD method shows significant promise for advancing molecular dynamics studies in various biological systems.