NCI-DOE Collaboration Paving Way for Large-Scale Computational Cancer Science

February 17, 2016, by Warren Kibbe, Ph.D.

Imagine the concentrated power of more than one million laptops working to screen a tumor sample from a patient against thousands of drugs and millions of drug combinations. At the end of this screening process, this mega-computer would help to identify a specific treatment with the greatest potential to combat that patient’s cancer.

NCI scientists, in collaboration with colleagues with the Department of Energy (DOE) Exascale Computing Initiative (ECI) and the National Strategic Computing Initiative (NSCI), have been hard at work for the past 14 months developing a plan to use this type of large-scale computing to influence cancer science and, ultimately, clinical treatment.

In a unique interagency initiative, leading scientists from NCI and several national DOE laboratories have been developing plans for a pilot collaboration that would help both agencies by significantly expanding NCI’s research capabilities and simultaneously advancing DOE’s efforts to develop energy efficient Exascale computing solutions.

Exascale computing refers to a system with the capability of making a billion billion calculations per second—also known as an exaFLOPS. Exascale is being studied as a means of improving the design of advanced materials, reverse engineering the human brain, and designing cost-effective renewable energy resources, among many other applications.

NCI and DOE leaders believe something with such massive power to interrogate and decipher big data could play an important role in cancer research.

As our plans have progressed, so too has the enthusiasm about the opportunities this collaboration presents to deliver important scientific insights, advance strategic computing, and offer new research opportunities for cancer investigators worldwide.

Consistent with the goals of NCI’s Precision Medicine Initiative, as well as the previously announced NSCI, the pilot activities emphasize three key areas:

• Developing new computational approaches to support research being done under NCI’s RAS Initiative
• Using large-scale computation to accelerate the development of patient-derived laboratory models of cancer
• Better understanding the impact of existing therapies outside of clinical trials, in real-world practice.

Considered individually, each of these pilot activities has significant potential to help advance our understanding of cancer. But when integrated with NCI’s research expertise and vast data resources and DOE’s computational modeling and supercomputing expertise, they can create opportunities and lead to insights that could pay tremendous dividends for years to come.

This unique collaboration will take advantage of NCI’s and DOE’s strengths, bringing together the remarkable power of supercomputing and the collective knowledge and wisdom of the cancer research community.

Through this collaboration, we’re beginning construction of the 21^st century digital launchpad needed to conduct next-generation cancer research and providing a framework for the many future missions that will produce critical innovations in cancer treatment. In doing so, we’re hopeful that we can build upon the tremendous progress that’s been achieved by the nation’s consistent commitment to advancing cancer research.