How AI Is Accelerating Drug Development for COVID-19
Drug researchers around the world are scrambling to develop vaccines and other treatments for COVID-19 in a desperate attempt to slow the spread of the disease. Thanks to AI’s capability to crunch massive amounts of data in a short period of time, drug development pipelines that would normally take over a decade are being compressed into a matter of months. But will it be enough?
Nobody knows whether the world’s drug researchers and pharmaceutical companies will be successful in finding one or more drugs that can treat COVID-19 patients – or better yet, give everybody immunity. The stakes are incredibly high, with nearly 4 million infections and over a quarter million deaths to date.
One of the companies in the middle of the search for COVID-19 treatments is IQVIA, a North Carolina firm that provides analytics services for pharmaceutical companies. The $11-billion business helps drug companies big and small accelerate their drug development pipelines by bringing not only its massive library of drug data, but also advanced analytics and AI capabilities (as well as cloud computing resources) to bear on its clients’ drug-development efforts.
AI is playing a critical role in IQVIA’s efforts to help its clients find treatments, therapeutics, and vaccines for COVID-19, according to Updesh Dosanjh, a technology solutions practice leader at the company.
“It’s really about taking the knowledge that our people have had and identifying the parts that are algorithmic,” Dosanjh tells Datanami in a recent interview. “In the past, you would put 40 PhDs into a room for 40 years and maybe come out with one drug. Now AI can help you identify 50,000 molecules an hour that you might be interested in.”
Once those 50,000 molecules are identified, another AI application helps the drug researchers narrow down the specific ones that they would be most interested in pursuing, Dosanjh says. “AI really helps me gets me to the right data and bring it to me and say, here’s the stuff that you should be looking at. You can ignore the 99% of information over here,” he says.
Other AI programs are used to determine how those molecules can help, and others still that can help determine if it’s safe (which is Dosanjh’s specific expertise). The company uses a range of machine learning models, including ones based on neural networks (which are particularly beneficial for analyzing textual data), to help clients find pharmaceutical substances that can solve specific problems.
“IQVIA is doing that analysis at each point in the information supply chain and essentially saying, Okay, AI can take that burden off you and bring you the results,” he says. “IQVIA supplies basically everything other than the final decision-making, which is what those companies should be focusing on. They shouldn’t be worrying about technology. They shouldn’t be worrying about AI algorithms. They should really be focused on saying what does the result look like, is it safe, and can I use it.”
AI Drug Pipelines
The novel coronavirus has spurred a global all-hands-on-deck effort to find treatments for COVID-19, the disease that it causes in people. The world’s top high performance computing (HPC) research outfits have stopped the work they’re doing and dedicated their supercomputer resources to finding cures (see the COVID-19 coverage of HPCwire, Datanami’s sister publication, for more information).
HPC-powered molecular modeling of novel microbes lives higher up the data supply chain than IQVIA. Instead, the company focuses the energy of its 67,000 employees on amassing all the knowledge that’s been created about microbial biology, disease, chemical pathways, and drug effects, and using that data to answer specific pharmaceutical questions in a rapid fashion. There are other companies in this field, but IQVIA arguably is the biggest and most advanced.
With the novel coronavirus ravaging the world, time is of the essence to find pharmaceutical treatments that work, and IQVIA has been quite busy helping drug companies do just that, Dosanjh says.
“You can imagine something like COVID where we’re trying lots of different things and trying them very rapidly, so instead of taking years and years, it’s now taking months,” he says. “The volume of data is not only is significant, but the speed at which that data is coming to organizations is rapidly different than what they’re used to in the past.”
Instead of finding novel cures to a novel illness, many researchers are combing through the extensive library of existing drugs to see which ones might be a good fit for treating COVID-19. This is an area that IQVIA’s extensive library of drug-related data and its AI models can really shine, Dosanjh says.
“I could have been the world’s greatest malaria expert for the last 30 years, and someone else says ‘Did you know this problem has been solved, and it kind of looks the same?’ That requires you to have a much broader data set than most companies have,” he says. “Sometimes you can be too stuck in your area to look at novel ways of handling data that might be relevant from something else.”
IQVIA generally runs about a billion queries per month against its massive data set on behalf of its clients. A considerable amount of that work is now dedicated to helping its clients – which span from the world’s biggest drug makers to the smallest research shops – explore possible COVID-19 treatments.
Dosanjh could not say specifically how IQVIA’s resources are being deployed against COVID-19. Considering how important it is to find treatments, it’s a substantial effort, he says. Despite the real that COVID-19 poses, existing research does not stop, he says.
“If we solve COVID and we don’t solve the next flu wave, then flu is going to be just as bad as COVID,” Dosanjh says. “They have to keep on doing that stuff. You’re already doing 15 products. Now you need to do 16, but the sixteenth is more like doing five products instead of just one.”
Billions of dollars are being spent to find treatments and a vaccine for COVID-19. Pharmaceutical companies are working feverishly to accelerate their research and development efforts in a bid to spare humanity from the medical and economic suffering from the global pandemic. Will we succeed? Nobody knows. But at least we know that IQVIA and its clients are bringing the latest AI techniques to bear on the problem, which gives us hope.
“Some of it comes down to luck and some of it comes down to just having the right data and having the tools to go and look at it,” Dosanjh says, “We’re confident the method we have is proven. Can I say that that means IQVIA will be the first to get there? No. All I can say is we’re bringing a set of proven technologies, maybe the only proven technology in the industry. I’m very confident that, if used in the right way by the right companies, it will be a significant help.”