Fast Parametrization of Martini3 Models for Fragments and Small Molecules

Area of Science:

• Computational chemistry and biophysics.
• Development of simulation methodologies.

Background:

• Coarse-grained molecular dynamics (CGMD) simulations, like those using the Martini 3 force field, simplify complex molecular systems for extended timescale and system size studies.
• Martini 3 offers enhanced flexibility with more bead types and sizes, facilitating the study of small molecules in phenomena like protein-ligand interactions and membrane permeation.
• Existing tools for automatic coarse-graining primarily support older Martini versions, creating a gap for Martini 3 compatible small molecule modeling.

Purpose of the Study:

• To introduce Auto-MartiniM3, an automated tool for generating coarse-grained models of small molecules specifically for the Martini 3 force field.
• To validate the accuracy and utility of Auto-MartiniM3-generated models through comparison with expert-designed models and experimental data.
• To assess the program's capability for high-throughput generation of coarse-grained models for diverse molecular systems.

Main Methods:

• Development and implementation of the Auto-MartiniM3 program for automated coarse-graining.
• Validation using 81 diverse small molecules from the Martini Database, comparing structural and thermodynamic properties.
• Assessment of model behavior through solute translocation and free energy calculations across lipid bilayers.
• Evaluation of complex molecule modeling, including caffeine binding to the adenosine A2A receptor.

Main Results:

• Auto-MartiniM3 successfully generated coarse-grained models for 81 small molecules, showing good agreement with expert-defined models.
• Simulations using Auto-MartiniM3 models accurately predicted solute translocation and free energy profiles across lipid bilayers.
• The tool demonstrated efficacy in modeling complex systems, such as caffeine-receptor interactions.
• Deployment on a large dataset confirmed the program's potential for fast, high-throughput model generation.

Conclusions:

• Auto-MartiniM3 provides an efficient and accurate solution for automating the coarse-graining of small molecules for the Martini 3 force field.
• The tool significantly lowers the barrier for utilizing Martini 3 in studies involving small molecules, enhancing research accessibility.
• Auto-MartiniM3 is poised to become a valuable resource for high-throughput molecular modeling in computational chemistry and biophysics.