Computationally Predicting and Characterizing the Immune Response to Viral Infections
Whether an individual mounts a strong response to SARS-Cov-2 or not depends, at least in part, on their genes. Specific genes code for the proteins on the surface of our cells that present viral protein fragments to the immune system. Killer T cells recognize these fragments and kill the infected cells. The immune response to COVID-19 hinges on whether the viral protein fragments bind into a groove in these cell-surface proteins -- like a key into a lock. The molecular biology is well understood. Whether a protein fragment binds or not is a question of 3D structure and simple atomic force calculations. The full proteome of the SARS-CoV-2 virus was published in March. However, no one has attempted to solve this problem for COVID-19 (or any other virus) because of the scale of the computation required.
The conventional approach is to simulate at the atomic level: the trajectories of all atoms in the system are determined by numerically solving Newton's equations of motion. Simulating a single binding event takes *days* of super-computing time. There are 21,000 variants of the relevant genes in the population, each coding for slightly different 3D structures of the cell-surface proteins. There are about 38,000 viral protein fragments from SARS-CoV-2. So the scale of the problem is to perform about 798 million such simulations.
We think that we can manage it. In joint work with the Mayo Clinic, we are develop custom software for this specific molecular problem, and we will deploy it at scale on cloud-computing infrastructure. With a targeted approach, we believe that we can turn 1 billion days of supercomputing time into 1 million minutes of cloud-computing time, which is manageable in terms of cost and complexity. Given reasonable cloud resources, the computation can complete in approximately one month.
Note that if we develop this ability, it will be transformative for COVID-19. The knowledge of which viral protein fragments are good targets for the immune system will enable vaccine development. If successful, the same computational infrastructure could be deployed in the future for other viruses. It could also be transformative in other contexts, for instance for treatments of cancer via immunotherapy as well as for the treatment of autoimmune diseases.
Marc Riedel is Associate Professor of Electrical and Computer Engineering at the University of Minnesota. From 2006 to 2011 he was Assistant Professor. He is also a member of the Graduate Faculty in Biomedical Informatics and Computational Biology. From 2004 to 2005, he was a lecturer in Computation and Neural Systems at Caltech. He has held positions at Marconi Canada, CAE Electronics, Toshiba, and Fujitsu Research Labs. He received his Ph.D. and his M.Sc. in Electrical Engineering at Caltech and his B.Eng. in Electrical Engineering with a Minor in Mathematics at McGill University. His Ph.D. dissertation titled "Cyclic Combinational Circuits" received the Charles H. Wilts Prize for the best doctoral research in Electrical Engineering at Caltech. His paper "The Synthesis of Cyclic Combinational Circuits" received the Best Paper Award at the Design Automation Conference. He is a recipient of the NSF CAREER Award.