They are racing against the clock, facing the monumental task of sifting through millions of potential treatments. They can combine them in the lab with the virus to see what happens, work that requires countless test tubes, chemicals and hours of work.
"It’s finding a needle in a haystack,” said Jerome Baudry, professor of biological sciences at University of Alabama at Huntsville. “It takes months and costs a fortune.”
That’s why Baudry and his team are bringing in the big guns. A supercomputer on loan from Hewlett Packard Enterprises will help Baudry search for potential coronavirus treatments in a fraction of the time.
“Instead of doing the tests in the test tube, we replicate that process in the computer,” Baudry said. “We can screen millions of compounds a day in a big computer and really focus on the most promising ones and reprocess them.”
His team will focus on a library of natural products found in living organisms. Other researchers are already focusing on existing drugs, but Baudry’s research will mine other chemicals that are readily available. They are partnering with the National Center for Natural Products Research at the University of Mississippi to identify and refine potential treatments.
If the group gets lucky, the computer will quickly identify a compound that can block the coronavirus from invading cells or replicating itself. At the very least, Baudry expects it to identify substances that show some ability to slow the virus and can be tweaked to become more effective.
They will initially run the computer at high speeds to identify the most promising compounds, then more slowly and methodically analyze those to determine which could prove most effective.
“We will use artificial intelligence to have a quick look at very large libraries to scale that down to maybe 20,000,” Baudry said. “The we take that top 20,000 and use other kinds of machine learning approaches to take a closer look at all of them.”
Hewlett Packard’s Cray Sentinel supercomputer is a powerful tool in the fight against coronavirus. Similar supercomputers have been used in drug discovery in the past by modeling chemical and physical reactions to predict how a given compound might act on viruses, Baudry said.
The problem is a difficult one. Baudry compared the coronavirus to a house with 14 doors. Researchers must find the right key – out of millions – to unlock one of those doors and get inside. Much of the world’s scientific establishment has pivoted to coronavirus research this year and is collaborating at a level Baudry said he has never seen.
“To have a chance to get something against the COVID virus, everyone has to try,” he said.
Baudry hopes his work will yield results quickly. The fight against coronavirus has some similarities to the Space Race, he said, and Huntsville already has partnerships in place to speech technological discovery.
“There is a philosophy here of let’s go to the moon quickly, not because it’s easy but because it’s difficult,” Baudry said. “I think we are the right place to make it happen.”
©2020 Alabama Media Group, Birmingham. Distributed by Tribune Content Agency, LLC.
