The world’s top supercomputers protect human health and safety in critical, never-before-seen ways, from predicting global weather patterns to examining structural safety during simulated earthquakes. Now City of Hope has vaulted into this league by setting up its own high-performance computing cluster, under the direction of Nagarajan Vaidehi, Ph.D., professor in the Division of Immunology.
Researchers use the new computing cluster to predict structures of membrane-bound proteins that are important drug targets for cancer therapy. They also employ it to design small-molecule drugs for cancer. Computational methods provide a rational, powerful, cost-effective and complementary approach to structure-based drug design, said Vaidehi.
The cluster’s power makes it an important component of the Developmental Cancer Therapeutics Program at City of Hope.
“We’re focused on developing computational methods to probe the structure, function and drug design for membrane proteins,” said Vaidehi. “These computational methods probe the three dimensional structure and function of certain membrane proteins such as G-protein coupled receptors that are difficult to study experimentally.”
Spencer Hall, Ph.D., assistant research scientist, and Allen Mao, research associate, helped to plan and build the computing cluster. It took nine months to design and construct the computing cluster, but only 10 days to get it fully operational once components were in place.
“I’d like to see a fully expanded cluster, with a continued expansion every year or so,” said Hall. “A consistent, sustained growth of the cluster allows us to increase our modeling capabilities and speed up the development of drug candidates.”
The cluster, housed in the Hilton Building, currently comprises 26 computers, with a total of 50 3.6 GHz Intel processors working together to process large jobs. It can complete a virtual screening of a compound in five seconds and can run through City of Hope’s entire library of 80,000 compounds in only seven days.
“These computational procedures not only make the drug-design process cost effective but also provide insight into the molecular mechanisms of the protein function,” said Vaidehi. “We’re not only gaining a more fundamental understanding of how proteins work, but also learning how drugs can be designed with less cross-reactivity.”
Vaidehi works in collaboration with Jack Shively, Ph.D., chair of immunology, on a chemokine receptor that plays an important role in more than 23 cancers. Her group also is a part of the STAT3 drug discovery team headed by Richard Jove, Ph.D., professor and chair of molecular medicine and co-director of the Developmental Cancer Therapeutics Program. She also conducts research into the design and optimization of nanotubes for drug delivery in conjunction with researchers at New York University.
Convenient access to the computing resources mean collaborations have the potential to grow, as well. System users can connect to the high-performance computing cluster from anywhere on campus.
“Eventually, I’d like to add an experimental base to our computational lab,” said Mao. “It allows us to verify our predictions and advance the computational methods.”
