Amid increasing global cases and threatening variants, a major gap remains in the global strategy to end the pandemic: effective therapies. Now, a Massachusetts startup powered by a wide array of HPC resources – including local resources from the Massachusetts Green High Performance Computing Center (MGHPCC) – has identified four promising drugs approved by the FDA that are effective as a COVID-therapy. 19 is supported by data from millions of people. patients.
“We have developed our own simulation suite,” explained Joy Alamgir, founder of the startup in question, called ARIScience, in an interview with HPCwire. “This was the genesis of ARIScience, where we wanted a new way of looking at compounds – in particular, the compounds that we want to interrupt – and a way of putting into quantum observations with classical simulation on the ligand side. … And we deliberately wanted to do this seamlessly and on the same development stack, which allows us to maintain it efficiently and also allows us to avoid any setup pain that we might have as we use different types of resources from calculation available to us. “
A few years later, of course, the landscape has changed somewhat. “Once COVID hit the United States around April of last year,” Alamgir said, “we redirected our efforts to see if there was an existing compound of all the FDA-approved drugs with which we could use our simulation platform to see if there was any. specific proteins of the coronavirus that we could interrupt. “
Alamgir wanted to simulate the structures of 1,513 FDA-approved drugs, then perform a free energy analysis against 11 key SARS-CoV-2 proteins to see which of these drug molecules had the best potential to disrupt those proteins.
Initially, Alamgir worked with ARIScience’s internal HPC cluster – a small set of “essentially three nodes” equipped with Intel processors dating back a few generations of hardware.
“Our internal HPC very quickly ran out of computing power,” said Alamgir, “at which point we contacted [John Goodhue, executive director of the] MGHPCC and a few other organizations to see if they could allow us to run it on their HPC platforms. Beyond the MGHPCC, Alamgir has received stipends at the Pittsburgh Supercomputing Center (PSC), the Texas Advanced Computing Center (TACC), and the universities of Maine and North Dakota.
None of the allocations were massive and Alamgir took advantage of the flexibility of the ARIScience platform to distribute the load across systems using Slurm. On each system, he estimated, the project used between six and 30 nodes at a time, with each simulation task taking a few hours and the 1,513 analyzes of a protein taking several days.
By the end of this whole simulation, Alamgir was left with 18 promising drugs approved by the FDA.
“How we were like, okay, great, we have these 18 results,” he said. “What do we do with them, don’t we? It’s not like we can do 18 clinical trials. “
Alamgir therefore decided to take the analysis a step further and check the results against real patient data. After “a lot of requests,” ARIScience gained access to data from the National COVID Cohort Collaborative (N3C) from the National Institutes of Health (NIH), making the company one of the first commercial entities to gain access to it. ‘huge data set. , which contains detailed – and yes, unidentified – data of 1.5 million patients.
Carefully controlling for demographics that might affect the results, Alamgir then used “very sophisticated statistical analysis” to examine the actual differences in mortality between patients who used any of these 18 drugs and those who did not. were not using.
The result: four drugs (amoxicillin, metformin, hydrochlorothiazide, and triamcinolone) that were each “statistically significantly associated with a reduction in COVID mortality of approximately 25%”. Additionally, although the sample size is smaller and data analysis continues, Alamgir shared that the combined effect of hydrochlorothiazide and metformin appeared to be even stronger. “The reduction in the death rate was greatest for patients who took both,” he said. “We have detected a 41% reduction in the odds of death from COVID.” (The other combined effects had not yet been assessed in a similar fashion.)
For now, Alamgir’s analysis is limited to mortality – the most crucial finding – and other patient data may be a bit more difficult to use. “We deliberately froze our analysis using a data release around mid-December 2020,” he said. “The main reason is that from the end of December the vaccination campaigns in the United States started and, depending on who was vaccinated or not, you are introducing additional unknowns into this statistical analysis.”
So now Alamgir and ARIScience are working with what they have: very promising results for four drugs approved by the FDA. With vaccines taking hold in the United States and critical peaks occurring elsewhere in the world, the company is looking to South America and South Asia to explore the possibility of a randomized clinical trial to strengthen still the effectiveness of drugs to prevent mortality from COVID.
For Goodhue and the MGHPCC, this is a familiar achievement, with the director describing the partnership between the MGHPCC and ARIScience as one of the centre’s many actions “at the cutting edge of research and at the forefront of societal impact ”. The centre’s previous accounts, he said, included the founders of Moderna and a range of successful startups. “They all started with us when they were at an early stage,” he said. “If I were to try to impress people, I would say we have been successful in helping a few companies on their way to exit through acquisition, which is a total of three quarters of a billion dollars.”
During the pandemic, the MGHPCC executed what Goodhue explains as a “mini version” of the COVID-19 HPC Consortium, providing computing resources to Massachusetts-based companies with research ideas to fight the virus. Right now, Goodhue said, the center is actively partnering with half a dozen companies.