Optimized Lennard-Jones Parameters for Druglike Small Molecules.

Area of Science:

• Computational Chemistry
• Molecular Modeling
• Drug Design

Background:

• Accurate molecular mechanical force fields are essential for computer-aided drug design (CADD).
• Existing methods like GAFF and CGenFF rely on predefined parameters, which can be inadequate for specific molecules.
• The previous GAAMP protocol optimized internal parameters and charges using quantum mechanics but had limitations in van der Waals interactions.

Purpose of the Study:

• To improve the accuracy of atomic models for small molecules in CADD.
• To address the limitations of van der Waals (Lennard-Jones) parameterization in the GAAMP protocol.
• To enhance the prediction of ligand physical properties, particularly hydration free energies.

Main Methods:

• Systematic optimization of Lennard-Jones 6-12 parameters using experimental data (neat liquid densities and enthalpies of vaporization) for 430 compounds.
• Rescaling of molecule-water van der Waals dispersion by a factor of 1.115 for optimal hydration free energy calculations.
• Validation of the optimized parameters through hydration free energy calculations.

Main Results:

• A new set of Lennard-Jones parameters was developed, improving the accuracy of molecular models.
• Optimal accuracy for hydration free energies was achieved with a specific rescaling factor for molecule-water interactions.
• The final optimized model demonstrated an average unsigned error of 0.79 kcal/mol in hydration free energies.

Conclusions:

• The developed Lennard-Jones parameters significantly enhance the accuracy of the GAAMP protocol for molecular modeling in CADD.
• The improved force field enables more reliable predictions of ligand physical properties, aiding in drug discovery.
• This work provides a more robust computational tool for characterizing protein-ligand interactions and other essential molecular properties.