Brookhaven National Laboratory Issues Update on Its Supercomputing Battle Against COVID-19

October 8, 2020

Oct. 8, 2020 — Researchers at the U.S. Department of Energy’s (DOE) Brookhaven National Laboratory are making progress on several fronts in the battle against COVID-19, the global pandemic sparked by the emergence of a novel coronavirus late last year. This work is part of a worldwide effort to understand the virus and the factors that affect its spread with the aim of devising treatments and other mitigation strategies.

“The urgency of this pandemic spurred researchers at Brookhaven and across the DOE complex to focus their efforts in areas where we could have real potential impact,” said John Hill, head of a working group coordinating Brookhaven Lab’s COVID-19 research response and director of the National Synchrotron Light Source II, one of the DOE Office of Science user facilities on the front lines in the search to discover treatments or vaccines. “It has been an intense time for all involved, but we have made real progress in understanding the details of this virus in a very short time.”

Even with most of the Lab in a “minimally safe” operating status from late March through the beginning of June, several beamlines at NSLS-II remained up and running to conduct critical COVID-19 research. Many other Brookhaven Lab scientists have been working from offsite on computational experiments essential to the drug- and vaccine-discovery effort. Lab scientists have also built epidemiological models to help predict how and where the disease could spread, and developed computational tools for tracking and extracting data from the enormous volume of virus-related research to help scientists streamline their search for solutions.

Much of the Lab’s work in this area is taking place in a collaborative way as part of the National Virtual Biotechnology Laboratory (NVBL), a consortium of DOE national laboratories focused on response to COVID-19, with funding provided by the Coronavirus CARES Act. This effort strategically draws on each lab’s core capabilities and unique facilities to maximize their impacts on the threats posed by COVID-19.

Learn more about Brookhaven’s efforts, including new experiments taking place as additional Brookhaven Lab research facilities reopen to continue the battle against COVID-19.

Drug and vaccine discovery

This slideshow of protein structures solved at NSLS-II includes images of potential inhibitor drugs bound to the SARS-CoV-2 main protease. Drugs that disable this enzyme could potentially block key steps of viral replication—like a wrench thrown into the viral machinery to jam up the works. Hover over image to reveal slideshow controls.

Discovering pharmaceutical drugs to thwart the coronavirus or vaccines to protect people from being infected requires being able to “see” how the virus and the cells it infects “talk” to one another at the molecular level. Brookhaven Lab has the tools and expertise to do just that.

For example, since the pandemic began, scientists have used the intense x-ray beams at NSLS-II to study a wide variety of virus and other COVID-19-related proteins and molecules. These include:

  • components of the “spike” protein the SARS-CoV-2 virus uses to attach to and break into cells
  • enzymes that help the virus hijack cellular machinery to make copies of itself and allow those copies to continue spreading infection
  • antibodies, known drugs, and small drug-like molecules that might disable the virus.

The x-rays at NSLS-II, among the brightest in the world, create atomic level pictures of these molecules to help scientists understand how they work—for example, which parts of the virus proteins bind with cell receptors in your body or activate and assist viral replication, which parts might be the best targets for drugs, and which drugs are most likely to be effective.

“The idea is to find molecular ‘wrenches’ to throw into the viral machinery and jam up the works—or identify parts of viral proteins that could be made into a vaccine to trigger an immune response,” said Sean McSweeney, head of the structural biology program at Brookhaven Lab. “Since March, we’ve worked with scientists from Brookhaven, academic institutions, and pharmaceutical companies to run studies on more than 5,500 samples. It is really quite a remarkable pivoting of many groups’ research to this new area; to be successful so quickly reflects the importance of our task,” he said.

Some of the samples scientists are studying come from best guesses based on years of biological research, including studies of related viruses and drugs that have been used to treat them. Other experiments are driven by computational studies that predict how various molecules will interact. These include models of how existing drugs (or molecules that could be made into drugs) might interfere with or disable viral proteins.

“Identifying potentially effective existing drugs that have already gone through safety studies may speed their deployment against this virus,” McSweeney said.

One promising target on SARS-CoV-2 is its main protease (MPro)—a protein that cuts newly replicated viral “polyproteins” into functional individual proteins, including those needed to assemble new infectious virus particles. Computational modeling studies at Brookhaven and elsewhere predicted that two drugs already approved for use in humans by the U.S. Food and Drug Administration (FDA) would bind MPro to block its function. So Brookhaven biologists crystallized this protease in complex with the potential inhibitor drugs, and studied the combined structures at NSLS-II. Their structures revealed in exquisite detail (1.5 angstrom resolution) that the drugs do indeed bind. The scientists are now working to improve the binding characteristics of these potential inhibitors.

Brookhaven researchers are also collaborating with scientists at the University of Texas Medical Branch’s Galveston Lab on studies of molecules that might disrupt the virus spike’s ability to mediate fusion with the host cell membrane. The most promising candidate drugs are being tested against live virus at the Galveston Lab, and these data will be necessary for a move to clinical trials.

For studies of larger protein complexes or proteins that are hard to crystallize for x-ray studies, scientists often turn to a technique called cryo-electron microscopy (cryo-EM). Brookhaven Lab was in the process of constructing a new cryo-EM center, funded in part by New York State, when the pandemic temporarily shut down the bulk of activity at the Lab. Because of its importance to the COVID-19 challenge, construction continued to bring this new Laboratory for BioMolecular Structure (LBMS) online by late July—months ahead of schedule.

Scientists from Brookhaven have been using the cryo-EM microscopes to study SARS-CoV-2 proteins that appear to play a role in some of the immune-system-triggered complications of COVID-19.

“These studies are aimed at understanding that process and could potentially point to ways to intervene,” said Qun Liu, the lead investigator in this work.

Complementary computational studies

As noted, computational simulations at Brookhaven and collaborating labs across the DOE complex and partner universities are playing an essential role in accelerating the search for antiviral drugs. Using structural information obtained from experimental studies, computational scientists can build computer-based models of viral proteins, simulate how those viral proteins interact with one another and/or human cells, and scan rapidly through libraries of billions of small drug-like molecules to see which might interfere with key viral functions.

“This computational work allows us to pre-screen and winnow down the many options to a much smaller set of highly probable candidate drug compounds, which can then be tested in experimental studies by our colleagues at the NSLS II and other similar facilities,” said Kerstin Kleese Van Dam, head of Brookhaven Lab’s Computational Science Initiative (CSI). “This is the ultimate, ultra-fast drug-discovery pipeline, leaving no stone unturned to help fight the virus.”

The idea is to test computationally how all known drug compounds and other possible molecular drug-like compounds might bind to weak points, or “pockets,” in the SARS-CoV-2 virus that have been identified by experiments.

“Think of it as trying out different puzzle pieces—just that you have not only many possible pieces, but more than one piece that fits into a given place, with some pieces fitting more easily and better than others,” said Kleese Van Dam. “While experimental studies are more accurate than our computational predictions, they do take much longer to accomplish. With so many possibilities—68 identified pockets and four billion possible drug compounds—we needed to find a faster way to home in on the real high potential candidates for drugs, antibodies, or vaccines.”

The Brookhaven team, together with NVBL partners, developed a scalable high-performance computing (HPC) and artificial intelligence (AI) infrastructure to help. This pipeline models and designs COVID-19 drug targets in three stages:

  • High-throughput ensemble docking studies identify small molecules that can bind to identified pockets on the virus.
  • AI-driven molecular dynamics simulations model atom-by-atom interactions in specific binding regions to understand mechanistic changes that take place during binding to elucidate how and how well a molecule binds to a pocket, which helps to identify the best leads.
  • Binding free-energy calculations optimize the most promising lead candidates.

The best candidate molecules identified through this process are then prioritized and handed over to experimental colleagues for validation.

“Each of these stages has been heavily optimized with physics-informed AI, replacing slow algorithms with fast machine-learning models, using AI to optimize the execution of both the physics-based models and the AI models, using AI to devise strategies to decide what to test next, and finally using AI to improve the overall pipeline performance and quality,” Kleese Van Dam said.

“With this approach, we are now able to scan 45 million compounds per hour!”

As an example of this approach in practice, CSI scientists from Brookhaven Lab and collaborators at Stony Brook University have built a complete computational model of the SARS-CoV-2 spike glycoprotein that simulates how it moves to an open (or up) position to bind to the “ACE2” receptor on human cells. The binding portion of this protein, they found, prefers a closed state, where it is buried in the spike glycan (sugar) shield, which appears to protect it from the body’s immune system.

But the team may have found an Achilles’ heel to block this protection—a pocket on the protein that changes shape during the down/up transition.

“After extensive screening, we’ve identified a number of molecules that could bind to this pocket, potentially locking the spike in an exposed ‘up’ state that would make it susceptible to the host’s immune response,” Kleese Van Dam said.

These computer-based findings are now being fed back to the experimental laboratories, where McSweeney’s group will use x-rays to study a fragment of the spike protein with and without the potential inhibitors.

Epidemiology modeling

Scientists at Brookhaven are also on the front lines of modeling how COVID-19 is spreading in communities. This effort draws on computational resources and expertise within CSI and also at the Lab’s Center for Functional Nanomaterials (CFN), another DOE Office of Science user facility. The aim is to identify the key factors that influence the infection rate to help communities decide when mitigation strategies such as lockdowns and school closures would be most effective—as well as when such measures might be safely lifted.

One model developed by CFN scientist Alexei Tkachenko in collaboration with CFN user Sergei Maslov of the University of Illinois at Urbana Champaign (UIUC) was used to advise officials in Illinois and has been described in a series of publications posted on preprint servers and submitted for peer review to scientific journals. These manuscripts include details of modeling second wave scenarios and factors that control herd immunity—a state reached when enough people in a community have developed immunity to make further outbreaks considerably less likely.

Most recently, the UIUC-led team has been working closely with university administrators to develop operating procedures as they reopen the university. There were some initial infection hot-spots—driven, the scientists say, by a combination of factors, including a small minority of students not complying with social-distancing guidelines. The university is refining its procedures, such as testing frequency, based in part on inputs from the models developed by the team, and the rate of positive tests is dropping again.

“We are involved in an unprecedented endeavor that aims to take control over the epidemic in a fairly large community (about 45,000 students, faculty, and staff),” Tkachenko said. “Our modeling is an integral part of a multicomponent system that also involves state-of-the-art testing, contact tracing, and an exposure-notification app. One of the challenges is that the model has to be perfected and calibrated based on real-time data, so the results of the modeling can then be used to adjust the policies.”

Aerosol modeling

Brookhaven also has experts in developing models of how tiny aerosol particles travel through the air. They normally focus on how atmospheric aerosols emitted by industry or natural sources such as sea spray affect the formation of clouds and climate. But in the time of COVID-19, their ability to track particle transport is just as relevant to understanding an airborne virus.

“We’re using aerosol microphysics simulations to study how respiratory aerosols—the tiny particles that come out of your mouth when you talk or exhale—disperse under different conditions,” said Laura Fierce, an associate scientist in the Environmental and Climate Sciences Department. “Our goal is to quantify how much virus you might be exposed to and your risk of infection based on how close you are to an infected person and how much time you spend with them talking and breathing the same air.”

Mask integrity studies

Given the evolving understanding that coronavirus particles may easily travel as aerosols, masks can play and important role in stopping the spread of the virus. But even from the early days of the pandemic, scientists knew that not all masks are alike. Furthermore, because mask shortages led many people to reuse what were once considered single-use commodities, there’s been an effort to understand how repeated use and various disinfection protocols might affect mask performance.

Amy Marschilok, Scientist and Energy Storage Division Manager in the Interdisciplinary Science Department at Brookhaven and an Associate Professor at Stony Brook University, has been using x-rays at the spectroscopy and imaging beamlines at NSLS-II, electron microscopy and x-ray photoelectroscopy at the CFN, and complementary measurements at ISB and SBU to examine the integrity of N95 and other mask materials following different types of disinfection protocols.

“We’ve been looking at the filter efficiency of mask materials using metal oxide nanoparticles to simulate the virus particles, and figuring out how to make our measurements sensitive enough to discriminate these from environmental contaminants,” she said. “We’re also studying materials from different layers of the mask filters to look for changes in fiber chemistry and orientation.”

Keeping track of thousands of studies

To avoid duplication of efforts and follow up on promising leads, scientists at Brookhaven and elsewhere must keep track of the latest developments in these research areas at laboratories all around the world. Brookhaven’s CSI group has come up with a tool to help. They’ve developed a database that catalogs thousands of scientific publications related to COVID-19 that have been published since the pandemic began, along with tools to find the most relevant articles for guiding next steps.

“We’ve cataloged more than 41,000 scientific papers since the pandemic began,” said CSI’s Kleese Van Dam. “Our search algorithms use ‘natural language processing,’ so scientists can simply pose questions to search through all that data.”

The searches provide access to publicly accessible publications and even those behind publishers’ pay walls for scientists at all of the DOE labs.

Brookhaven has also provided infrastructure for a COVID-19 archive repository to compile experimental and computational results, papers, reports, presentations, and software to share vetted results with the research community. Scientists from across the DOE complex can submit items into this long-term repository, which will aid current efforts and support research into responses to future pandemics.

“Our goal throughout this effort,” concluded COVID-19 research working group chair John Hill, “has been to combine our strengths, at Brookhaven and across the DOE labs, to tackle this enormous challenge as fast as we can. We have some unique resources, including computational resources and state-of-the-art x-ray and electron tools, and the world’s best scientists working on this problem. We are working with our colleagues across the country and across the world night and day. Together we will succeed.”

Computational and structural biology studies aimed at developing therapeutic drugs, mask integrity studies, and aerosol modeling are all supported by the National Virtual Biotechnology Laboratory (NVBL), a consortium of DOE national laboratories focused on response to COVID-19, with funding provided by the Coronavirus CARES Act, distributed by the DOE Office of Science (BER). Operations at NSLS-II and CFN are supported by the DOE Office of Science (BES). The publication tracking database was funded by Brookhaven’s CSI. The UIUC team is supported by the University of Illinois System Office, the Office of the Vice-Chancellor for Research and Innovation, the Grainger College of Engineering, and the Department of Physics at the University of Illinois at Urbana-Champaign.

About BNL

Brookhaven National Laboratory is supported by the U.S. Department of Energy’s Office of Science. The Office of Science is the single largest supporter of basic research in the physical sciences in the United States and is working to address some of the most pressing challenges of our time. For more information, visit https://www.energy.gov/science/


Source: Brookhaven National Laboratory

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industry updates delivered to you every week!

MLPerf Inference 4.0 Results Showcase GenAI; Nvidia Still Dominates

March 28, 2024

There were no startling surprises in the latest MLPerf Inference benchmark (4.0) results released yesterday. Two new workloads — Llama 2 and Stable Diffusion XL — were added to the benchmark suite as MLPerf continues Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing power it brings to artificial intelligence.  Nvidia's DGX Read more…

Call for Participation in Workshop on Potential NSF CISE Quantum Initiative

March 26, 2024

Editor’s Note: Next month there will be a workshop to discuss what a quantum initiative led by NSF’s Computer, Information Science and Engineering (CISE) directorate could entail. The details are posted below in a Ca Read more…

Waseda U. Researchers Reports New Quantum Algorithm for Speeding Optimization

March 25, 2024

Optimization problems cover a wide range of applications and are often cited as good candidates for quantum computing. However, the execution time for constrained combinatorial optimization applications on quantum device Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at the network layer threatens to make bigger and brawnier pro Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HBM3E memory as well as the the ability to train 1 trillion pa Read more…

MLPerf Inference 4.0 Results Showcase GenAI; Nvidia Still Dominates

March 28, 2024

There were no startling surprises in the latest MLPerf Inference benchmark (4.0) results released yesterday. Two new workloads — Llama 2 and Stable Diffusion Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing po Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HB Read more…

Nvidia Looks to Accelerate GenAI Adoption with NIM

March 19, 2024

Today at the GPU Technology Conference, Nvidia launched a new offering aimed at helping customers quickly deploy their generative AI applications in a secure, s Read more…

The Generative AI Future Is Now, Nvidia’s Huang Says

March 19, 2024

We are in the early days of a transformative shift in how business gets done thanks to the advent of generative AI, according to Nvidia CEO and cofounder Jensen Read more…

Nvidia’s New Blackwell GPU Can Train AI Models with Trillions of Parameters

March 18, 2024

Nvidia's latest and fastest GPU, codenamed Blackwell, is here and will underpin the company's AI plans this year. The chip offers performance improvements from Read more…

Nvidia Showcases Quantum Cloud, Expanding Quantum Portfolio at GTC24

March 18, 2024

Nvidia’s barrage of quantum news at GTC24 this week includes new products, signature collaborations, and a new Nvidia Quantum Cloud for quantum developers. Wh Read more…

Alibaba Shuts Down its Quantum Computing Effort

November 30, 2023

In case you missed it, China’s e-commerce giant Alibaba has shut down its quantum computing research effort. It’s not entirely clear what drove the change. Read more…

Nvidia H100: Are 550,000 GPUs Enough for This Year?

August 17, 2023

The GPU Squeeze continues to place a premium on Nvidia H100 GPUs. In a recent Financial Times article, Nvidia reports that it expects to ship 550,000 of its lat Read more…

Shutterstock 1285747942

AMD’s Horsepower-packed MI300X GPU Beats Nvidia’s Upcoming H200

December 7, 2023

AMD and Nvidia are locked in an AI performance battle – much like the gaming GPU performance clash the companies have waged for decades. AMD has claimed it Read more…

DoD Takes a Long View of Quantum Computing

December 19, 2023

Given the large sums tied to expensive weapon systems – think $100-million-plus per F-35 fighter – it’s easy to forget the U.S. Department of Defense is a Read more…

Synopsys Eats Ansys: Does HPC Get Indigestion?

February 8, 2024

Recently, it was announced that Synopsys is buying HPC tool developer Ansys. Started in Pittsburgh, Pa., in 1970 as Swanson Analysis Systems, Inc. (SASI) by John Swanson (and eventually renamed), Ansys serves the CAE (Computer Aided Engineering)/multiphysics engineering simulation market. Read more…

Choosing the Right GPU for LLM Inference and Training

December 11, 2023

Accelerating the training and inference processes of deep learning models is crucial for unleashing their true potential and NVIDIA GPUs have emerged as a game- Read more…

Intel’s Server and PC Chip Development Will Blur After 2025

January 15, 2024

Intel's dealing with much more than chip rivals breathing down its neck; it is simultaneously integrating a bevy of new technologies such as chiplets, artificia Read more…

Baidu Exits Quantum, Closely Following Alibaba’s Earlier Move

January 5, 2024

Reuters reported this week that Baidu, China’s giant e-commerce and services provider, is exiting the quantum computing development arena. Reuters reported � Read more…

Leading Solution Providers

Contributors

Comparing NVIDIA A100 and NVIDIA L40S: Which GPU is Ideal for AI and Graphics-Intensive Workloads?

October 30, 2023

With long lead times for the NVIDIA H100 and A100 GPUs, many organizations are looking at the new NVIDIA L40S GPU, which it’s a new GPU optimized for AI and g Read more…

Shutterstock 1179408610

Google Addresses the Mysteries of Its Hypercomputer 

December 28, 2023

When Google launched its Hypercomputer earlier this month (December 2023), the first reaction was, "Say what?" It turns out that the Hypercomputer is Google's t Read more…

AMD MI3000A

How AMD May Get Across the CUDA Moat

October 5, 2023

When discussing GenAI, the term "GPU" almost always enters the conversation and the topic often moves toward performance and access. Interestingly, the word "GPU" is assumed to mean "Nvidia" products. (As an aside, the popular Nvidia hardware used in GenAI are not technically... Read more…

Shutterstock 1606064203

Meta’s Zuckerberg Puts Its AI Future in the Hands of 600,000 GPUs

January 25, 2024

In under two minutes, Meta's CEO, Mark Zuckerberg, laid out the company's AI plans, which included a plan to build an artificial intelligence system with the eq Read more…

Google Introduces ‘Hypercomputer’ to Its AI Infrastructure

December 11, 2023

Google ran out of monikers to describe its new AI system released on December 7. Supercomputer perhaps wasn't an apt description, so it settled on Hypercomputer Read more…

China Is All In on a RISC-V Future

January 8, 2024

The state of RISC-V in China was discussed in a recent report released by the Jamestown Foundation, a Washington, D.C.-based think tank. The report, entitled "E Read more…

Intel Won’t Have a Xeon Max Chip with New Emerald Rapids CPU

December 14, 2023

As expected, Intel officially announced its 5th generation Xeon server chips codenamed Emerald Rapids at an event in New York City, where the focus was really o Read more…

IBM Quantum Summit: Two New QPUs, Upgraded Qiskit, 10-year Roadmap and More

December 4, 2023

IBM kicks off its annual Quantum Summit today and will announce a broad range of advances including its much-anticipated 1121-qubit Condor QPU, a smaller 133-qu Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire