NVIDIA Awards $400k to Trailblazers in Cancer Research
News Nov 24, 2016
Up close and personal with cancer. That describes how two research teams are approaching their fight, using accelerated computing, to battle the disease at the most basic levels of human biology.
Each group will receive $200,000 from the NVIDIA Foundation to further research that could lead to new and more targeted 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.
A group of NVIDIA employees, with the support of researchers at the National Cancer Institute, chose the recipients from among nearly 20 proposals submitted from around the world.
One team, led by Seungchan Kim at the Translational Genomics Research Institute (TGen), based in Phoenix, Ariz., aims to understand why some cancer cells respond to treatments and some don’t. This could make it possible to treat different cells with different medications.
The other team, led by Andrés Cisneros at the University of North Texas, is searching for mutations that cause changes in the proteins that repair damage to DNA. These mutations may indicate the presence of cancer.
Fighting Cancer, Cell by Cell
Research by Kim and his team at TGen could lead to precise cancer therapies that target specific cells within each patient’s tumor.
To do so, they created a GPU-accelerated statistical analysis tool that allows them to examine in precise detail how the cells’ DNA controls protein production and how those proteins interact with each other and with other molecules. Using the tool, the researchers are able to identify differences among groups of cells in the same tumor.
This could lead to personalized treatments for cancer patients that treat different parts of the tumor with different medicines.
“There may be a subpopulation of cells that respond only to compound A and another subpopulation of cells that respond only to compound B,” Kim said.
The tool, called EDDY-GPU (EDDY is short for “evaluation of differential dependency”) allows researchers to analyze the vast amount of data from thousands of tumor cells quickly enough to make a difference for patients, Kim said. In contrast, an earlier CPU-based version of EDDY worked on 4,754 samples for two months and never finished.
“If you have a cancer patient with their life on the line, you don’t want to wait for results,” Kim said.
He plans to use the Compute the Cure funds to improve his algorithms so they work even faster.
Kim and his team are also planning a trial with the University of California at San Francisco. Using tissue samples from two brain cancer patients, they’ll perform detailed analysis of the patients’ tumors, propose therapies and investigate these further using patient-derived models.
DNA is frequently damaged by UV light, the carcinogens in cigarette smoke and many other substances. Fortunately, the body manufactures DNA repair proteins that normally take care of the problem.
But sometimes the DNA repair proteins themselves break or mutate. Cisneros and his team at the University of North Texas are combing through vast datasets from the National Institutes of Health and others to find mutations on DNA repair proteins that are related to cancer.
After identifying the mutations, the researchers use GPU-accelerated computer simulation to figure out how these change the DNA repair protein and its ability to do its job.
“If we know there are mutations that affect these proteins and they’re related to cancer, then researchers can use this info to try to fix the proteins or come up with drugs or other therapies to treat the disease,” Cisneros said.
Cisneros’ larger goal is to find more biomarkers that point to an elevated risk for particular types of cancer. Already the team has identified several mutations related to different types of cancer, including a biomarker related to prostate cancer in African Americans.
Scientists currently use genetic data to determine risk for certain types of cancer and to determine treatment. For example, physicians can test for mutations linked to a high risk for breast cancer and testing for receptors for estrogen or progesterone helps doctors determine which treatments are likely to be most effective.
By identifying more biomarkers, Cisneros could provide new diagnostic tools for more types of cancer. In addition, his detailed analyses of mutations on DNA repair proteins could help scientists develop personalized, genetic therapies to fight the disease.
Story from NVIDIA Foundation. Please note: The content above may have been edited to ensure it is in keeping with Technology Networks’ style and length guidelines.
An artificial intelligence (AI) approach based on deep learning convolutional neural network (CNN) could identify nuanced mammographic imaging features specific for recalled but benign (false-positive) mammograms and distinguish such mammograms from those identified as malignant or negative.