Neural potentials of proteins extrapolate beyond training data

We evaluate neural network (NN) coarse-grained force fields compared to traditional CG molecular mechanics force fields. We conclude NN force fields are able to extrapolate and sample from unseen regions of the free energy surface free energy surfaces when trained with limited data. Our results come from 88 NN force fields trained on different combinations of clustered free energy surfaces from four protein mapped trajectories. We used total variation similarity as our metric to assess the agreement between free energy surfaces of force fields and the mapped atomistic simulations. Additionally, force matching error was found to only be weakly correlated with a force field's ability to reconstruct the correct free energy surface. These conclusions support the common hypothesis that constructing force fields on one region of the protein free energy surface can indeed extrapolate to unexplored regions.