How Supercomputing and Advanced X-Rays Helped the Government Fight COVID-19
Years of basic research and speedy, strategic coordination helped to catalyze new and needed treatments.
Government-backed scientists spent countless hours in 2020 applying advanced research and sophisticated technological capabilities to help the United States better grasp and effectively fight COVID-19.
Much of their efforts built on prior work that started unfolding long before the pandemic even posed a threat, but modern, real-world impacts hastened by the global emergency are now beginning to come to light.
For instance, genetic mutations within and helping strengthen five potential vaccines—including those developed by Pfizer/BioNTech and Moderna—hinged upon more than a decade of research advanced by the Energy Department’s Advanced Photon Source, or APS.
“The development of the COVID-19 vaccines may seem like an overnight success story, but it was based on years of research, some of it at publicly funded facilities like the APS at Argonne National Laboratory. Knowledge gained at the APS during these years helped make COVID-19 vaccines from Pfizer and Moderna more effective,” Robert Fischetti, group leader of Argonne’s X-ray Science Division and Life Sciences Advisor to the APS Director recently told Nextgov via email. “This research was done without any idea of how important it would be for us in 2020.”
Here’s a look at how the U.S. government brought powerful technologies like the APS and supercomputers to the fight against the modern pandemic.
Computing vs. COVID-19
America boasts some of the most powerful high-performance supercomputing resources on the planet, several of which were quickly pivoted and pooled to hone in on the novel coronavirus almost as soon as it surfaced. Such advanced systems can compute heaps of data much faster and with more accuracy than traditional machines—and in the pandemic, speed was of the essence. Ad hoc and federally coordinated supercomputing efforts enabled researchers to run cycles spanning a diverse array of simulations, programs and models meant to strengthen understanding of COVID-19, and develop treatment options like antiviral drugs and vaccines.
“It’s important to recognize that COVID-19 research has been an Argonne priority in 2020,” Fischetti said.
Viruses reproduce themselves within cells of living hosts—like humans, animals and beyond. Early into the pandemic, researchers at the Energy Department’s Illinois-based lab leveraged supercomputing capabilities to conduct host analysis and ultimately determine how various individuals respond to COVID-19. Around that time, they also began creating and refining epidemiological models to simulate how the virus spreads across populations. That work helped pave the way for what would become CityCOVID—the most detailed simulation of COVID-19 spread in the Chicago area—steered by Argonne scientists and supported by its Theta supercomputer. Among other efforts, and in conjunction with the APS, officials from Argonne’s supercomputing centers have also used artificial intelligence methods to identify vaccine candidates and treatments via supercomputer-powered simulation.
“Argonne computer scientists have identified at least 50 molecules that could be effective treatments against the virus, and as part of the National Virtual Biotechnology Laboratory, a collaboration between national labs, have zeroed in on a drug compound that attacks a key part of the virus,” Fischetti noted. “Argonne scientists are now using those machine learning techniques to analyze the new mutation of the virus that has been seen in the United Kingdom.”
Other labs and federal entities with next-level technological and computational assets also pivoted to prioritize confronting the virus in the early months of 2020.
The Pentagon’s High Performance Computing Modernization Program was quick to provide capacity to help conduct virtual drug screenings of potential COVID-19 vaccine candidates in the spring and also launched computational fluid dynamics studies to support the military’s assessment of safely airlifting patients.
Around that time, researchers at Sandia National Lab also shifted to use genetic sequencing tools and CRISPR-based technologies—which help thoroughly probe the most micro happenings inside of cells—to ultimately genetically engineer antiviral countermeasures to stop the current and other future outbreaks. While Lawrence Livermore National Lab’s Corona supercomputing system (which was named prior, for the total solar eclipse in 2017) was previously being used for unclassified science applications, it was also turned to underpin research to virtually screen, design and validate antibody candidates for SARS-CoV-2, the virus that causes COVID-19, and help shed light on possible antiviral compounds. That system, as well as another at Argonne saw upgrades in support of their drug-driving research.
And in the weeks after COVID-19 really set in as a threat, Oak Ridge National Lab granted researchers emergency computation time to sift through a database of existing drug compounds for combinations that might prevent cell infection of COVID-19 and perform simulations with the Summit supercomputer in support of finding a cure. That machine was then considered the world’s fastest supercomputer, but in the summer, Japan’s Fugaku surpassed it on the Top 500 list of global supercomputers. The non-U.S. supercomputer remains in the leading spot.
Still, both powerful systems are now among many others that make up America’s COVID-19 High Performance Computing Consortium.
Speedily spearheaded by the White House Office of Science and Technology Policy, Energy Department, and tech-giant IBM in late March, the initiative was meant to accelerate discoveries against the novel coronavirus by catalyzing unprecedented new access to some of the U.S.’ weighty supercomputing resources. Through the still-ongoing effort, participating national laboratories, agencies, companies and academic institutions volunteer to share free compute time and assets that scientists can apply to tap into from anywhere for COVID-19 research.
New members have steadily joined the consortium, pushing its initial computing capacity of 330 petaflops to about 600 petaflops, currently. And more than 90 in-the-works and active research projects are now harnessing the massive computing capabilities for a wide range of investigations. They include studies that use artificial intelligence and machine learning to advance new drug discoveries, one that taps a NASA-run supercomputer to define risk groups for disease associated with the novel coronavirus, another supporting the design of certain devices for coronavirus patients—and many more. In November, officials involved in the public-private partnership revealed it’s moved to its “second phase,” which will prioritize research projects that could likely advance patient outcomes and treatments in the next half-year. The shift came in the wake of positive vaccine trials led by Moderna and Pfizer.
OSTP went on to release a request for information in late December regarding lessons learned through the consortium. The notice articulates officials’ aims to see the initiative blossom into a broader computing reserve to boost the U.S.’ readiness for future threats.
“The prompt, successful, and nimble deployment of computational resources (including expertise) via the COVID-19 High-Performance Computing Consortium has demonstrated its essential role in the nation's response to emergencies,” it reads. “This backdrop has led to the conceptualization of a National Strategic Computing Reserve, comprising a coalition of experts and resource providers that could be mobilized quickly to provide critical computational resources (including compute, software, data, and technical expertise) in times of urgent need.”
X-Ray Beaming Breakthroughs
The Energy Department’s APS offers capabilities like those of an extreme microscope—it uses high-energy X-rays that enable operators to see through materials and biological structures with almost unmatched detail. Knowing how viruses are shaped proves crucial when designing means to fight them. So, a primary contribution of the APS against the modern killer virus is that it offers an intensely bright X-ray beam that can deliver very precise pictures of the molecules that make it up.
“Designing vaccines and treatments is like making a key, and the virus is like a lock,” Fischetti said. “The better you can see the grooves and bumps of the lock, the better your key will work.”
More than 5,000 researchers tap into the APS for science explorations each year and most studying viruses via the machine now use its facility remotely. They can control the whole thing from their home institutions, Fischetti said, noting that some will mail in samples and then use robots to load those samples onto an X-ray beamline.
“Much what we know about the structure of this virus comes from APS data,” he explained.
Fischetti said scientists began pinpointing structures of the proteins that make up the SARS-CoV-2 virus that causes COVID-19 right when its genetic sequence was released in January. “More than 80 research groups from across the country have logged more than 10,000 hours of time at the APS to study SARS-CoV-2, determining more than 100 structures of the virus’s proteins” since then, he noted.
The rollout of COVID-19 vaccines marks the speediest development in history, but Fischetti emphasized that it wouldn’t have come so quick had it not been founded on more than a decade of previous research.
“Our response to the COVID-19 pandemic was tied to basic science research conducted over the past two decades,” he said. “It stands to reason that our response to the next pandemic will be heavily influenced by the science we support and fund in the years before it arrives.”
One team over the years had turned to the APS to help guide the design of new vaccines and found in the pandemic that some of the same techniques they’d previously developed also work well against SARS-CoV-2. Like many other viruses, that new coronavirus attacks cells through a protein that juts off of the surface of the virus and latches onto human cells.
“The key discovery the team made was that these proteins are more vulnerable before they latch onto the cell. The researchers designed a way to train the body to attack the virus before it infects cells,” Fischetti said. “They used the APS to refine that technique into what is now in several of the COVID-19 vaccines.”
Federal funding via the Coronavirus Aid, Relief, and Economic Security or CARES Act enabled the APS to operate additional hours over the summer so therapeutics-hunting research could continue, which was critical. But to Fischetti it’s important to consider that some of the most tangible outcomes were years-in-the-making.
“This is just the latest example of the need to ensure a steady stream of funding for basic science research,” he said. “We don’t know what today’s seemingly incremental discoveries may lead to in the future.”