Computational Science and Engineering
Through the invention of new computational techniques, the faculty in this group focus on scientific discovery in the natural sciences and engineering novel systems. Meet some of the faculty active in this group below.
Professor of Aerospace and Mechanical Engineering
“Computing is fundamental to making discoveries in much of the physical sciences. Certainly, this is true of biophysics and materials physics, which are disciplines that interest me.”
There are mathematical descriptions underpinning these fields that have a generality stretching far beyond a specific phenomenon. As an example, the dynamics governing the migration, proliferation and transition to cancerous behavior of cells are shared by swarms of animals, populations of humans and communities, and even the way opinions form and spread through society. Their mathematical complexity means that computation is often
the only way to tease out the emergent phenomena in these systems. In turn, physics inspires the design of his computational methods.
Using computational and scientific machine learning methods, he and his team have discovered that when cells with cancerous potential are signaled to by certain non-cancerous stromal cells, they can organize themselves, forming patterns, and transitioning to an aggressive, rapidly migrating and proliferating behavior. Yet wound healing by the same cancer cells is mainly by them going on a disorganized, random walk. These results have important implications for understanding how tumors metastasize and spread to distant organs.
USC Associates Chair in Natural Sciences and Professor of Chemistry
“In chemistry and physics, computing is an integral part of research. Quantum chemistry calculations connect experimental observations with a detailed atomistic picture of underlying processes. Such calculations are essential for deriving insight and advancing our understanding of molecules and materials.”
Krylov is developing new quantum chemistry approaches and computer codes for the description of electronically excited and open-shell species, including electronically metastable states. These methods are implemented in the Q-Chem software package, which is one of the leading codes for molecular simulations. Her group is also using these computer codes to investigate the role of radicals and electronically excited species in combustion, solar energy, bioimaging, spectroscopy, and quantum information science.
Her important accomplishments include: development of spin-flip approach to strong correlation and extending coupled-cluster theory to spectroscopy modeling in high-energy and high-intensity regimes.
Professor of Quantitative and Computational Biology, Chemistry and Pharmacology and Pharmaceutical Sciences
“My field of research concerns biochemistry and structural biology and their applications to the discovery of new therapeutic drugs. Computing, both physics-based molecular modeling and data-driven deep learning/AI, are the key to modern drug discovery, dramatically reducing its time and cost requirements and improving resulting therapies.”
His lab is purely computational, and works in close collaboration with experimental labs that test his team’s predictions. Most notably, Katritch’s group has developed V-SYNTHES - a new computational approach to screening chemical spaces of billions of compounds for new drug candidates. This technology, now with a deep learning accelerator module, has become a core platform for the CNT3D center, where his team collaborates across USC and around the globe to facilitate early drug discovery for clinically relevant targets.