Students, Faculty Head Back to School, but a Few will be Staying On
August 26, 2017
Their posters rolled up after a lively Aug. 3 presentation session, nearly all of the students and guest faculty working in the Berkeley Lab Computing Sciences area are now back home, preparing for the new school year.
But not all of them. Rafael Zamora and Tom Corcoran from Hood College in Maryland have had their stays extended as they are applying deep learning techniques to classifying specific protein structures that could lead to more effective drugs for fighting cancer.
Both Zamora and Corcoran graduated with degrees in computer science last June and when their professor, Xianlin Liu, said he had a visiting faculty position at the lab for the summer, they both agreed to join him.
“Professor Liu was a user of NSF’s Blue Waters and tries to get his students interested in HPC,” Zamora said. “This opportunity at Berkeley has been a great experience – the people are very friendly and the weather and location are great.”
Zamora also appreciated the presentations and activities organized by both the Computing Sciences Summer Student program and the lab’s Workforce Development and Education office.
“The summer project at LBNL was a great experience for me and my students,” said Prof. Liu. “The Visiting Faculty Program is a rare opportunity for someone like me who teaches at a small college when we have a brief period of time to concentrate on research.”
The team from Hood College worked with lab scientist Silvia Crivelli to develop deep learning techniques to study Ras proteins, which play a critical a role in cellular proliferation. More than 30 percent of all human cancers, including 95 percent of pancreatic cancers and 45 percent of colorectal cancers, are driven by mutations of the RAS genes, which create mutant forms of the Ras proteins. These variants are ‘switched on’, which may cause cancer cells to reproduce uncontrollably. If the Ras proteins and the various Ras mutations can be fully characterized, that could help pharmaceutical firms design drugs that would specifically dock with the mutated proteins to fight the cancer. Their effort is focused on the 3D structure of the proteins because the structure determines how they function.
“We proposed a novel algorithm to allow 3D structural data to be analyzed by convolutional neural networks (CNN),” Prof. Liu said. “Our preliminary work in the summer demonstrated the concept using protein model classification tests. Any high dimensional data can be processed for CNN using our algorithm, which utilizes locality property of space-filling curves.”
“We used deep learning to classify the Ras proteins and that worked well, so we stepped up to a harder classification problem – differentiating between H-Ras and K-Ras proteins,” Zamora said. “The whole approach of deep learning is take an unbiased look at the structure, which may help find features that others overlooked. We don’t tell the program what to look for, but train it to ‘learn’ on its own.”
As the summer program wound down, Zamora presented one of the 55 posters at the Computing Sciences student poster session. Both Zamora and Corcoran are taking a short vacation, and will then return to the lab to continue their work for a few weeks, with the hope of extending it even longer.
“Hood College is a small liberal arts college and we weren’t familiar with it, but they came here and did very good work,” Crivelli said. “All of the participants were excited by the opportunity to work on real-life problems and many of them are motivated to continue working with HPC.”
Liu, Zamora and Corcoran were part of a group of eight faculty-student teams doing research in the Computational Research Division. Two teams were supported by funds from the Department of Energy’s Office of Advanced Scientific Computing Research supporting Computing Sciences’ Sustainable Research Pathways Program. Six teams were supported through the Department of Energy’s Visiting Faculty Program, which brings faculty members and students to national labs to work on research problems. In early December, Computing Sciences will sponsor a workshop to begin recruiting participants for next summer.
“Our goal is to make connections with faculty we might not otherwise connect with and to contribute to our diversity,” said CRD Director David Brown, who worked with Mary Ann Leung of the Sustainable Horizons Institute to develop the program, now in its second year. “With the undergrads, we hope to inspire them to go to graduate school and then later they will hopefully consider working at a DOE lab. Their posters were really good and the students were very enthusiastic when talking about what they did here.”
Other faculty-student teams working in CRD were:
- Prof. Sally Ellingson and student William Derek Jones from the University of Kentucky, Lexington, who worked on computational prediction of adverse drug reactions with Wibe de Jong in the Computational Chemistry, Materials & Climate Group. Read more about their work.
- Prof. Tiffani Holmes and student Courtney Small of Fort Valley State University, Georgia, also worked with de Jong on drug-biomolecule interactions and quantitative structure activity relationship (QSAR) study of some novel cathinone derivatives. Watch a short video of Prof. Holmes talking about her experience.
- Prof Wai Lau and students Xiaoxue (Amy) Hu, Elise Bishoff, and Kelsey Sabu from Seattle Pacific University in Seattle worked on load balancing strategies for AMR applications with Ann Almgren in the Center for Computational Sciences and Engineering.
- Prof. Bogdan Czjedo and students Catherine Spooner and Casey Lorenzen from Fayetteville State University, North Carolina, worked on developing new machine learning techniques and tools in protein folding research with Silvia Crivelli in computational biology.
- Prof. Ovais Khan and student Joshua Gaston from Tuskegee University, Alabama, worked on numerical investigation of oblique shock waves and shock reflection with Dan Graves in the Applied Numerical Algorithms Group (ANAG).
- Prof. Alina Lazar and student Zackary Harnett from Youngstown State University, Ohio, worked on efficient clustering algorithms for real-time streaming data with Alex Sim and John Wu in the Scientific Data Management Group.
- Prof. Jae Ryu and student Jason Stewart Walters from the University of Idaho worked on enhancing drought outlooks for western water management in a changing climate with Michael Wehner in the Computational Chemistry, Materials & Climate Group.
And the following students were also at the lab as part of the program:
- Joe Umhoefer and Diana Gonzalez from Oregon State University worked with Dan Martin and Hans Johansen in ANAG;
- Conrad Czjedo from the University of North Carolina, Chapel Hill, and Itzhel Dimas of Cal Poly Pomona worked with Silvia Crivelli in computational biology.
About Computing Sciences at Berkeley Lab
The Computing Sciences Area at Lawrence Berkeley National Laboratory provides the computing and networking resources and expertise critical to advancing Department of Energy Office of Science research missions: developing new energy sources, improving energy efficiency, developing new materials, and increasing our understanding of ourselves, our world, and our universe.
Founded in 1931 on the belief that the biggest scientific challenges are best addressed by teams, Lawrence Berkeley National Laboratory and its scientists have been recognized with 13 Nobel Prizes. Today, Berkeley Lab researchers develop sustainable energy and environmental solutions, create useful new materials, advance the frontiers of computing, and probe the mysteries of life, matter, and the universe. Scientists from around the world rely on the Lab’s facilities for their own discovery science. Berkeley Lab is a multiprogram national laboratory, managed by the University of California for the U.S. Department of Energy’s Office of Science.
DOE’s Office of Science is the single largest supporter of basic research in the physical sciences in the United States, and is working to address some of the most pressing challenges of our time. For more information, please visit energy.gov/science.