How do you screen billions of drug compounds to find the right one? Connect a research team at the University of Alberta with a supercomputer 2,700 km away in Ontario using Canada’s high-speed national research and education network (NREN).
Leveraging this powerful infrastructure, Dr. Michael Houghton and colleagues at the University of Alberta’s Li Ka Shing Applied Virology Institute are speeding up the time it takes for life-saving drugs to be identified from months or years to weeks.
The research team of virologists is using one of the fastest supercomputers in Canada to model molecular dynamics and quickly sift through billions of candidate molecules to find ones with the right therapeutic action, and minimal side effects.
SOSCIP’s 8-rack IBM Blue Gene/Q supercomputer, which includes a ½-rack owned by Li Ka Shing Institute of Virology, provides the computational horsepower. Its combined peak theoretical performance is 840 teraflops—a blazingly fast processing speed for computational drug discovery. The supercomputer is based at SciNet, a supercomputing consortium located at the University of Toronto.
Even though the research team is in Edmonton, a dedicated network connection provides secure access to run the superfast simulations. Canada’s federated NREN model sees 12 provincial and territorial partners – including in this instance ORION in Ontario and Cybera in Alberta – partner with CANARIE.
“The process and results have been fantastic,” says Houghton, who emphasizes that drug discovery can be a laborious process, sometimes taking years to identify potential drugs. Using supercomputers to do computational drug discovery, Houghton believes, is the future.
“Not only does this use of computational science provide a faster and more powerful way to discover drugs, but it gives us the power to be competitive with the big pharmaceutical companies, who are still using traditional biological-screening methods.”
An early vote of confidence came when the Li Ka Shing Institute researchers investigated a new hepatitis C drug, which was created by a large pharmaceutical company but had unexpectedly caused cardiac failure in participants during clinical trials.
One of the research fellows at the Li Ka Shing Institute theorized that a specific ion channel in the heart was being blocked by the drug. The research team modelled components of the ion channel using the IBM Blue Gene systems, which demonstrated that the drug would, indeed, block it. The prediction was later confirmed by subsequent biological screening.
“If these data had been available before the clinical trials, they would have never happened,” says Houghton.
Due to the complexity of finding the right molecular shape, computational drug discovery requires enormous processing speeds on powerful computers, such as the SOSCIP IBM Blue Gene/Q. Houghton and his colleagues, Dr. Jack Tuszynski and Dr. Lorne Tyrrell, initially envisioned installing their own newly-bought Blue Gene in Edmonton at the University of Alberta, but were soon introduced to the benefits of using Canada’s National Research and Education Network.
SOSCIP, a research and development consortium that pairs academic and industry research with advanced computing tools to fuel innovation, approached the Li Ka Shing researchers to suggest integrating their IBM Blue Gene with SOSCIP’s, to give both organizations more combined computing power. Access to the high-speed network has made the partnership run seamlessly.
The Li Ka Shing Institute research team is now working on several anti-viral and anti-cancer drug design projects, including one that boosts the immune system to help the body fight cancer cells. The computational and biology teams, which are led by Dr. Khaled Barakat, have identified several molecules that inhibit immune system checkpoints.
The research shows promise for boosting the immune system in fighting off cancer, and the teams are developing candidate molecules into pharmaceutical drugs.
Houghton estimates that a lead drug that can be prepared for clinical testing in cancer patients could be identified within the next two years, thanks to the relationship between Alberta and Ontario. Despite the fact that these two provinces are thousands of kilometres apart, distance disappears and discovery accelerates when supported by Canada’s NREN.
For more information please contact our contributor(s):