Replica-exchange enveloping distribution sampling – a robust method to estimate multiple free-energy differences from a single simulation

The driving force for all (bio)molecular processes is the change of free energy. The estimation of free-energy differences from computer simulations remains, however, a challenging task. Rigorous methods to estimate free-energy differences that rely on conformational sampling from molecular dynamics (MD) or Monte Carlo (MC) simulations belong to the most accurate but also computationally expensive approaches. The majority of these methods are only suited to calculate free-energy differences between two end states. The efficiency could be substantially increased if multiple free-energy differences could be calculated from a single simulation. This is in principle possible with the method “enveloping distribution sampling” (EDS), but its application has been hampered by the difficulties encountered in the estimation of optimal parameters for the EDS reference state. To resolve this, we have recently developed a generalised form of the replica-exchange enveloping distribution sampling (RE-EDS) approach. By combining EDS with replica exchange, the complexity of the parameter-choice problem can be substantially reduced and the full potential of EDS can be exhausted. We have shown the successful application of the methodology to calculate 36 free-energy differences between nine small organic molecules with diverse substituents from a single MD simulation.

Full paper:
Sidler, D., Schwaninger, A., Riniker, S. (2016): Replica exchange enveloping distribution sampling (RE-EDS): A robust method to estimate multiple free-energy differences from a single simulation, AIP The Journal of Chemical Physics, 145, 154114

Prof. Dr. Sereina Riniker