Google Keynote: Fighting COVID-19 with Google Cloud and VirtualFlow: Harnessing the Scale of Cloud to Accelerate Discovery of COVID-19 Therapeutics
Using VirtualFlow in the Google Cloud, we screened over 1 billion chemical compounds as candidate treatments against the SARS-CoV-2 virus. Utilizing the scale of the Cloud, we were able to rapidly apply 100 million CPU hours of compute and accelerate the screening process by two orders of magnitude compared to our on-premise resources. Targeting 17 proteins important for the virus’ replication, we have carried out simulations of ~50 billion docking instances against 40 target sites. This allowed us to quickly narrow the chemical catalog down to a few thousand candidates, some with the potential for broader pan-coronavirus effectiveness.