Application of molecular dynamics simulations in molecular property prediction II: diffusion coefficient

Area of Science:

• Computational chemistry
• Molecular dynamics simulations
• Physical chemistry

Background:

• The Assisted Model Building with Energy Refinement (AMBER) force field is widely used for molecular simulations.
• Accurate prediction of dynamic properties, such as diffusion coefficients, is crucial for understanding liquid behavior.

Purpose of the Study:

• To evaluate the performance of the AMBER force field in predicting diffusion coefficients of various solutes in different liquid environments.
• To develop and test an efficient sampling strategy for calculating diffusion coefficients.

Main Methods:

• Molecular dynamics (MD) simulations were performed using the AMBER force field.
• Diffusion coefficients were calculated for 17 solvents, organic compounds, and proteins in aqueous and nonaqueous solutions.
• An efficient sampling strategy involving averaging mean square displacements from multiple short MD simulations was employed.

Main Results:

• Diffusion coefficients for organic solutes in aqueous solutions were well predicted with low average unsigned errors (0.137 × 10(-5) cm(-2) s(-1)) and root mean square errors (0.171 × 10(-5) cm(-2) s(-1)).
• Good correlations (R(2) > 0.78) were observed between predicted and experimental diffusion coefficients for organic solvents, proteins in aqueous solutions, and organic compounds in nonaqueous solutions.
• Temperature-dependent behaviors of TIP3P water, dimethyl sulfoxide, and cyclohexane were investigated.

Conclusions:

• The AMBER force field, combined with the proposed efficient sampling strategy, is effective for predicting diffusion coefficients of solutes at infinite dilution.
• The methodology shows promise for studying dynamic properties of liquids and solutions across various chemical and biological systems.