Update May 23, 2018: The British Journal of Cancer has published a paper describing how Compute the Cure grant recipient John Quackenbush and his research team developed a new way of analyzing genomic data that could lead to treatments that attack many types of cancer.
Researchers commonly search vast datasets of thousands of cancer patients’ genomic information to identify mutations linked to cancer. Quackenbush found that it was far more effective to group data according to biological systems that control functions like immune system response, DNA repair and cell division. By doing so, he identified characteristics shared by multiple cancer varieties known as subtypes.
“We tend to think about cancer as being of the liver, kidney, breast or other part of the body, but when we dig into the mutations, we understand that mutations are more universal,” Quackenbush said. “Without the NVIDIA grant, we never would have taken on this challenge and or have made our discoveries.”
Read the paper, Cancer subtype identification using somatic mutation data. Our original story follows.
Our employee-led philanthropic arm is awarding $400,000 in grants to two teams of researchers who are using innovative computing methods to turn old ideas about cancer research inside out.
Each group will receive $200,000 from the NVIDIA Foundation for its work in using Big Data to collect, analyze and distribute ever larger quantities of information in ways that are reshaping their investigative fields.
One team, led by John Quackenbush, professor of biostatistics and computational biology at Boston’s Dana-Farber Cancer Institute, is crunching genomic data obtained from thousands of cancer patients to uncover genomic patterns that could lead to new cancer treatments.
The other, led by Vijay Pande, professor of chemistry, structural biology and computer science at Stanford University, has invited people to contribute their unused home-computing power to aid in the team’s search for new breast cancer treatments.
The two grants are part of the NVIDIA Foundation’s Compute the Cure effort, which supports projects that use parallel computing technology to yield breakthroughs in cancer treatment and diagnostics. (For background, check out “NVIDIA Foundation, National Cancer Institute to Distribute Up to $2M to Cancer Researchers.”)
A team of our employees, with the support of researchers at the National Cancer Institute, chose the recipients from among nearly two dozen proposals submitted from around the world.
Falling Cost of Genomics Data
They were drawn to the work of Quackenbush and his team for the way their work reimagines the search for cancer treatments. With the costs of genomic data rapidly falling, the Dana-Farber Cancer Institute launched a program to collect tumor genomic data from more than 10,000 patients each year, along with detailed clinical data on how each individual has responded to therapy.
As the collection grows, Quackenbush wants to use GPUs to speed up the search for data patterns that could identify genetic subtypes that might ultimately be linked to new treatments.
This effort will involve, in part, a search for genetic similarities among different types of cancer. So a treatment used for colon cancer might be applied to a cancer originating elsewhere in the body but sharing similar patterns of mutation.
“We may have a drug today targeting a specific mutation in colon cancer, but which has never been tried in breast cancer,” Quackenbush says. “If we can identify patterns across diseases, we can more quickly identify potential new therapies.”
Folding@home: Harnessing a Sea of GPUs
Vijay Pande’s work is similarly bold. He’s not just looking for a cure for cancer. The one-time gamer – he admits to a weakness for Starcraft – is inviting computer enthusiasts worldwide to join the search. Started in 2000, Folding@home harnesses the power of tens of thousands of GPUs to form the world’s largest distributed supercomputer.
“This is an opportunity for other people to become part of the project and really make a difference,” Pande says.
The Folding@home software program simulates how proteins in the body assemble themselves. When this “folding” process misfires, it can lead to diseases. Researchers hope to create treatments for these “misfolds.” Data from the project has already helped produce more than 100 research papers.
With the NVIDIA Foundation’s help, Pande wants to push Folding@home further by combining cutting-edge molecular simulation and machine learning techniques with Folding@home’s user network to better understand and predict patient-specific tumor mutations. The goal: more targeted and effective treatments – first for breast cancer and, eventually, for other types of cancer.