The cloud offers the capability for massive scale-up of computational resources. While many see the value in faster turnaround of existing tools, we also see the possibility for applying more accurate and computationally more expensive methods. We have leveraged the cloud to perform large-scale virtual screens using tens of thousands of processors to accelerate the screening process, where we included partial receptor flexibility through the use of a structural ensemble. In addition, we have deployed free energy perturbation (FEP+) on GPUs in the cloud in order to accurately predict binding potency of thousands of molecules. Calculations that previously would have taken months or years can now be turned around in a matter of days, which fit into the timelines of drug discovery projects and allow for greater impact.

Using the cloud for massive scale-up of computationally intensive physics-based simulations

253rd National Meeting (2017)

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