Machine Learning/Molecular Mechanics (ML/MM) methods for the simulation of chemical reactions

Quantum Mechanics/Molecular Mechanics (QM/MM) methods have been instrumental in simulating biological systems such as enzyme catalyzed reactions. However, the efficiency of QM/MM simulations is greatly limited by the cost of the QM calculation. Machine learning potentials offer a solution to this computational bottleneck. Machine learning potentials are capable of predicting energies and forces at the quantum mechanical level of accuracy, but at a fraction of the cost. In this work we combine machine learning with a QM/MM approach to develop a ML/MM method in order to study solvated reactions and enzyme catalysis.