NNP/MM: Accelerating Molecular Dynamics Simulations with Machine Learning Potentials and Molecular Mechanics

We are pleased to announce the release of an optimized implementation of the NNP/MM method, which combines a neural network potential (NNP) and molecular mechanics (MM). Our upgraded implementation significantly enhances simulation speed, approximately quintupling its efficiency, thereby rendering it a valuable tool for conducting biomolecular simulations.

To showcase the capabilities of our implementation of NNP/MM, we conducted molecular dynamics simulations on various protein–ligand complexes and metadynamics simulations on a ligand. Remarkably, we achieved a combined sampling of 1 μs for each complex, marking the longest simulations ever reported for this class of simulations.

You can read more about our implementation here.

‍