Achieving Reproducibility and Replicability of Molecular Dynamics and Monte Carlo Simulations Using the Molecular...

Area of Science:

• Computational Chemistry
• Materials Science
• Chemical Engineering

Background:

• Molecular simulations (MD/MC) predict properties but face reproducibility challenges.
• Previous studies show discrepancies in simulation data, even with common force fields.
• Complexity in molecular systems can lead to inherent simulation errors.

Purpose of the Study:

• To introduce the Molecular Simulation Design Framework (MoSDeF) for transparent and reproducible simulations.
• To demonstrate MoSDeF's capability in standardizing molecular model implementation across different simulation engines.
• To validate MoSDeF's effectiveness in achieving consistent density predictions with increasing model complexity.

Main Methods:

• Utilized the MoSDeF workflow package to initialize molecular simulation setups.
• Applied MoSDeF to standardize the implementation of molecular models for various simulation engines.
• Conducted simulations with varying system complexity to assess density prediction consistency.

Main Results:

• MoSDeF initialization leads to consistent system density predictions.
• The framework ensures reliable results even when dealing with more complex molecular models.
• Standardization via MoSDeF mitigates variability in simulation outcomes.

Conclusions:

• MoSDeF facilitates TRUE (Transparent, Reproducible, Usable-by-others, and Extensible) molecular simulation studies.
• The framework enhances the reliability and consistency of thermophysical property predictions.
• MoSDeF is a valuable tool for advancing the accessibility and accuracy of molecular simulations.