Data-driven generation of perturbation networks for relative binding free energy calculations

A data-driven approach for predicting networks for affinity calculations offers a new route for automated molecular simulations in drug discovery.