A multidisciplinary team including members of the Frederick National Laboratory has been recognized with the “Best Paper” award at SC19, a prestigious international conference for high-performance computing, networking, storage, and analysis. Of the 344 papers submitted to SC19, only 87 were accepted, and just one was named Best Paper.
The research team spent more than two years working on their project, the second pilot arm of an initiative known as the Joint Design of Advanced Computing Solutions for Cancer (JDACS4C). This collaboration, which began in 2016, hopes to accelerate cancer research by using cutting-edge high-performance computing.
Titled “Massively Parallel Infrastructure for Adaptive Multiscale Simulations: Modeling RAS Initiation Pathway for Cancer,” the team’s paper describes a method of predictively modeling the interactions between RAS proteins and the lipids that make up cell membranes. RAS signaling only becomes activated when the proteins are bound to lipid bilayers; therefore, understanding this interaction has wide-ranging implications for therapeutic intervention.
One of the paper’s most innovative aspects is the development of a multiscale model that can provide high-resolution detail at a molecular level over long periods of time, an extremely challenging and resource-intense endeavor. The researchers also found that the model (called a Multiscale Machine-Learning Modeling Infrastructure, or MuMMI) is scalable for use on some of the most powerful computers in the world, including Lawrence Livermore National Laboratory’s Sierra and Oak Ridge National Laboratory’s Summit computers.
The NCI RAS Initiative, led by the Frederick National Laboratory, tested and validated predictions from the computer models to ensure their accuracy compared to experiment-derived biological interactions. This data was fed back into the simulations to improve the model. That iterative process helps to refine both high-powered computing and RAS research.
“The application of state-of-the-art molecular dynamic simulations to an intractable cancer biology problem is yielding unprecedented insights into cancer biology that may lead to new strategies for targeting KRAS-driven cancers,” said Dwight Nissley, Ph.D., who was a co-author on the study and is the director of the national lab’s Cancer Research Technology Program and head of the RAS Initiative.
The project is entering its fourth year and will next include the RAS effector protein RAF1 to better understand the activation of signaling that drives many cancers. The team will continue to use Department of Energy computers, including Summit at Oak Ridge National Laboratory, which is currently the world’s most powerful supercomputer.
By Chris Worthington, staff writer; photo courtesy of Lawrence Livermore National Laboratory