Search papers, labs, and topics across Lattice.
This paper introduces Garnet, a graph neural network that predicts force field parameters for molecular dynamics simulations directly from molecular structure, using continuous atom typing. Garnet was trained from scratch on quantum mechanical data, condensed phase simulations, and protein NMR data, without relying on existing force field parameters. The resulting force field achieves performance comparable to existing force fields on a range of systems, including small molecules, folded and disordered proteins, and protein complexes, while also demonstrating competitive performance in relative binding free energy predictions.
A graph neural network can learn accurate force field parameters from scratch, rivaling manually-developed force fields and opening avenues for automated force field discovery.
Force fields for molecular dynamics are usually developed manually, limiting their transferability and making systematic exploration of functional forms challenging. We developed a graph neural network that assigns all force field parameters for diverse molecules using continuous atom typing. The freely-available model, called Garnet, was trained on quantum mechanical, condensed phase and protein nuclear magnetic resonance data without the use of existing parameters. The resulting force field shows comparable performance to current force fields on small molecules, folded proteins, protein complexes and disordered proteins. It shows similar results to popular approaches for relative binding free energy predictions across a range of targets. Assessing different functional forms shows that the double exponential potential is a flexible and accurate alternative to the Lennard-Jones potential. Garnet provides a platform for automated, reproducible force field discovery that brings the benefits of machine learning to classical force fields.
