Case Study: Harvard Medical School speeds COVID-19 drug discovery with Google for Education

Harvard team conducts ultra-large virtual drug screenings on Google Cloud to speed up COVID-19 research

Harvard Medical School was looking for a way to speed up COVID-19 drug discovery by screening enormous libraries of chemical compounds much faster than traditional wet-lab methods allow. Working with Google for Education, the team used VirtualFlow on Google Cloud to run ultra-large virtual drug screenings against viral and human protein targets.

Google for Education provided cloud computing support and Google Cloud research credits that enabled Harvard Medical School to scale to 80,000 cores and screen more than one billion compounds in five days. The team completed 35 billion docking instances using 75 million CPU hours, identified top-scoring candidates for further study, and planned to publish results in Google’s Public Datasets to help advance broader COVID-19 research.

Harvard Medical School

Christoph Gorgulla

Postdoctoral Fellow