By Dave Bukey
Two groups of Computation Institute researchers at the U.S. Department of Energy's (DOE) Argonne National Laboratory earned special awards from the office of the U.S. Secretary of Energy for addressing the global health challenges of Ebola and cancer.
For answering urgent questions about potential Ebola outbreaks in the U.S., Argonne scientists Chick Macal, Jonathan Ozik and Nick Collier earned the DOE Secretary’s Appreciation Award. Also recognized, for helping predict precise therapies for specific cancer patients, was the team of Rick Stevens, Tom Brettin and Ravi Madduri.
“I am thrilled that, once again, our scientists and engineers have received the Department of Energy’s most prestigious awards,” said Paul Kearns, Argonne’s interim director. “These six researchers deserve exceptional recognition for taking meaningful steps toward protecting our nation from disease.”
Both teams are harnessing Argonne’s Mira supercomputer, located at Argonne Leadership Computing Facility, to approach public health dilemmas in new and promising ways.
As the Ebola virus swept through West Africa in 2014, data scientist Macal and his team in Argonne’s Global Security Sciences division joined DOE’s Ebola Task Force to halt the international health crisis.
“The DOE asked us to model what might happen if Ebola came to the U.S.,” said Macal. “We dropped everything and got to work.”
Macal, Ozik and Collier have spent years refining sophisticated models that can simulate human responses to crises; they could come quickly to realistic answers.
The group uses an approach that predicts the behavior of individual “agents” – representing people, households, organizations, etc. – and how they interact. With agent-based modeling, Macal can see how entire populations may evolve and identify tipping points that affect their fate in countless scenarios, including an Ebola outbreak.
Macal’s team had extensive computer models of how populations of large U.S. cities may handle a bacterial infection such as Methicillin-resistant Staphylococcus aureus, or MRSA. To adapt the models for Ebola, they visited Chicago-area hospitals and incorporated Ebola-related procedures into their models.
“This allowed us to estimate the healthcare resources required – space, staff and how to dispose of medical waste,” Macal said.
By January 2015, as the Ebola epidemic peaked, Macal, Ozik and Collier had a playbook detailing how a city like Chicago might respond to – and suppress – the spread of such a disease.
Collaborating with other national laboratories, Argonne improved “the robustness and quality of analytic products used to inform critical decisions throughout the outbreak,” wrote Richard Hatchett, former acting director of the Biomedical Advanced Research and Development Authority at the Department of Health and Human Services, in support of the awards.
The team has since generalized their models to apply to a wide range of infectious diseases and created the Chicago Social Interaction Model (chiSIM), an agent-based model of more general social interactions.
Meanwhile, Argonne’s Stevens, Brettin and Madduri introduced a similarly radical approach to cancer research. In January 2016, the Cancer Moonshot Task Force was launched to renew governmental focus on cancer research.
With two decades of experience in computational biology, Stevens jumped at the opportunity. “The immediate goal was to accelerate research on cancer and to pioneer the notion of predictive medicine,” he said.
Stevens and his team took the award for their role in the partnership between DOE and the National Cancer Institute, which was named the Joint Design of Advanced Computing Solutions for Cancer.
The project, in which Ozik and Collier also were involved, brings the full capacity of advanced computing and machine learning to cancer, examining the disease from three perspectives: molecular, patient and population.
Stevens’ group focuses on patient challenges, interpreting pre-clinical models that can predict how specific patients and tumors may respond to different types of drugs. (For more, see “Cancer’s big data problem”).
“More than 50 percent of patients with tumors don’t respond to chemotherapy,” said Stevens. “This is a huge challenge. We are trying to streamline the building of models that would give us a recommendation – for this patient, try these set of drugs in this order.”
Stevens and his colleagues are starting the second of three years devoted to discovering computational clues to beat cancer.
The Ebola modeling research was supported by the Argonne’s Laboratory Directed Research and Development program. The ongoing cancer research is funded by U.S. Department of Energy’s Exascale Computing Project, the National Institutes of Health and the National Nuclear Security Administration. The Argonne Leadership Computing Facility is a DOE Office of Science User Facility.
Originally published at the Argonne National Laboratory website.