An Australian research group is helping identify existing drugs to treat COVID-19 through the use of powerful data and computing services.
Professor Alan Mark and his team at the University of Queensland are using a molecular modelling tool, called the Automated Topology Builder (ATB), to determine the various shapes and atomic interactions in drug compounds that could be repurposed to fight COVID-19.
The team is rushing to incorporate all possible forms of more than 7,300 pharmaceutically active compounds, which have already passed Phase II clinical trials and thus are safe for human use, into their database.
The sophisticated computational process has been made possible through the use of Australian data services and computing power, including the Australian Research Data Commons’ (ARDC) Nectar Research Cloud, which provides a self-service structure for storing, accessing and analysing data. The ATB also draws on Pawsey’s Nimbus cloud service, the ARDC supported QRIScloud and the Gadi supercomputer at the National Computational Infrastructure (NCI).
Researchers around the world are using the ATB’s data in their efforts to find COVID-19 treatments.
“Worldwide, thousands of groups are trying to decide which drugs will bind to the virus. The more accurately the molecules can be described, the more useful these predictions are,” Alan said.
Assessing the potential of some of these compounds, which involve larger molecules, requires access to special computing capacity, which few researchers have.
“Only one fourth to one third of drugs have less than 50 atoms. The rest require high level calculations, which requires much larger computers. Some calculations take days to weeks even on Australia’s top facilities,” Alan said.
Until recently, the ATB could only process compounds with 50 atoms or less at a high level. However, due to the COVID-19 crisis, his team now has additional resources for their very timely work.
“The project would not be possible without the generous support of both very large and smaller scale computer facilities in Australia and around the world,” Alan said.
“Nectar Research Cloud, for example, allows us to farm out the work over multiple nodes and lets us run them in parallel.”
Nimbus is being used to process larger compounds. Its flexibility allows the team to run around 20 compounds asynchronously across multiple nodes without blocking the pipeline.
The ATB now includes 6,800 Phase II compounds in its database. The remaining compounds and alternative biologically relevant forms will be available soon.
The database was created 10 years ago.
“It started with just the computers in my lab,” said Alan. “Then we branched out to using Nectar Research Cloud; we went from processing 100’s to 10,000’s molecules per year. It’s now growing faster and faster.”
The ATB repository is one of the largest pre-calculated molecular structure and parameter databases in the world. It contains over 430,000 molecules, which are freely accessed by more than 12,600 users worldwide.