Harnessing Deep Learning for Optimization of Lennard-Jones Parameters for the Polarizable Classical Drude Oscillator Force Field

The outcomes of computational chemistry andbiology research, including drug design, are significantly influencedby the underlying forcefield (FF) used in molecular simulations.While improved FF accuracy may be achieved via inclusion ofexplicit treatment of electronic polarization, such an extensionmust be accompanied by optimization of van der Waals (vdW)interactions, in the context of the Lennard-Jones (LJ) formalism inthe present study. This is particularly challenging due to theextensive nature of chemical space combined with the correlatednature of LJ parameters. To address this challenge, a deep learning(DL)-based parametrization framework is developed, allowing forsampling of wide ranges of LJ parameters targeting experimentalcondensed phase thermodynamic properties. The present workutilizes this framework to develop the LJ parameters for atoms associated with four distinct groups covering 10 different atom types.Final parameter selection was facilitated by quantum mechanical data on rare-gas interactions with the training set molecules. Thechosen parameters were then validated through experimental hydration free energies and condensed phase thermodynamicproperties of validation set molecules to confirm transferability. The ultimate outcome of utilizing this framework is a set of LJparameters in the context of the polarizable Drude FF, which demonstrated improvement in the reproduction of both experimentalpure solvent and crystal properties and hydration free energies of the molecules compared to the additive CHARMM General FF(CGenFF) including the ability of the Drude FF to accurately reproduce both experimental pure solvent properties and hydrationfree energies. The study also shows how correlations between difference in the reproduction of condensed phase data betweenmodel compounds may be used to direct the selection of new atom types and training set molecules during FF development.

Automated summary

The outcomes of computational chemistry and biology research are greatly influenced by the choice of forcefield used in molecular simulations. This study develops a deep learning-based framework for optimizing van der Waals interactions in molecular simulations. The resulting LJ parameters (interactions between atoms) are validated for their performance in reproducing condensed phase thermodynamic properties and demonstrate improved accuracy in reproducing solvent and crystal properties.