Computational tools are often used in drug discovery. In particular, virtual screening can reduce the costs and time investments associated with early-stage hit identification.1 Standard virtual-screening methods rely on a single rigid protein receptor into which flexible small molecules are docked. The advantage of this approach is its speed, but it does not consider the conformational richness of the receptor and so may miss some small-molecule binders. In contrast, ensemble docking leverages an ensemble of protein conformations, increasing the chances of finding ligands.2 Brute-force molecular dynamics (MD) simulations are often used to generate these protein ensembles, but such simulations often get trapped in local minima and can take too long to thoroughly sample the full conformational landscape.3 In this work, we use weighted ensemble (WE) simulations4,5, an enhanced sampling technique that draws on many simulation trajectories running in parallel, to accelerate the sampling of protein pocket conformations. This novel WE method, called SubPEx, uses a predefined progress coordinate (i.e., a specific representation of the configuration space) to focus computational power on pocket sampling. We applied different progress coordinates to three test systems and observed an increase in sampling in all three cases.

Sub-Pocket EXplorer (SubPEx): Leveraging weighted ensemble simulations to enhance the conformational search of binding-pocket conformations

ACS Spring 2022

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