Computational approaches are widely employed throughout the drug discovery process. Computational tools such as virtual screening, docking and QSAR play a key role in both identifying hit molecules and designing strategies to optimise those hits. Free energy calculations are now widely used by many companies to predict the binding of a specific candidate molecule to a target protein binding site. These methods can be used to rank compounds, prioritise synthesis and accelerate the optimisation process.<br/>A technical limitation within the current standard of computational methods for calculating ligand binding free energies is the effect of inadequate sampling within such simulations on the accuracy of the predicted binding free energies. Sampling different ligand binding pathways entails surmounting free energy barriers associated with important conformational changes within the ligand and associated active site rearrangements. Recent work has also indicated the effect the initial structure chosen for intermediate alchemical states can have upon the accuracy of the free energy calculation for ligand binding.<br/>A method of enhanced sampling is presented herein for performing relative binding free energy calculations using alchemical perturbations. The protocol combines the ability of Sequential Monte Carlo (SMC) targeting specific ligand and side chain torsions to propose an optimally minimum number of intermediate states for a given perturbation and the long-timescale sampling achievable by simulated tempering (ST) over the proposed intermediate states. Further optimisations of the perturbation protocol during the simulation yields a fully adaptive form of simulated tempering (FAST) that can be used to obtain accurate estimates of binding free energies in an automated and system-independent manner.<br/>Here, the ability of FAST to provide accurate relative binding free energies for protein-ligand binding cases associated with sampling rare events will be presented. The adaptive method also yields a series of diagnostics that indicate the efficiency in sampling binding pathways given the current set of intermediate states. The use of these diagnostics to inform the further development of FAST to tackle more complex conformational pathways of ligand binding will also be discussed.

Diagnostics for the evaluation of relative binding free energy calculations using Fully Adaptive Simulated Tempering (FAST)

Fall 2022 abstracts, posters, and presentations will launch in mid-September.

ACS Fall 2022

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.