COVID-19, our current coronavirus, is far from the first coronavirus. Most people likely know another coronavirus, albeit by a different name: SARS. From 2002 to 2004, SARS – standing for “severe acute respiratory syndrome,” and officially known as SARS-CoV – ran its course as a pandemic, killing nearly 800 people. Now, we face its much more infectious, much deadlier cousin: SARS-CoV-2. Researchers at the University of Arkansas leveraged supercomputing resources at the Texas Advanced Computing Center (TACC) to explore the similarities and differences between these two coronaviruses, yielding important contributions for our understanding of the new pandemic.
The University of Arkansas research team, led by computational chemist Mahmoud Moradi, created detailed 3D simulations of the molecular dynamics of the spike proteins on both coronaviruses. These proteins, which the viruses use to bind to human cells, are crucial to their ability to infect human hosts.
The simulations were run on two supercomputers at TACC. The first, Frontera, comprises 8,008 compute nodes equipped with Intel Xeon CPUs, 192 GB of memory, 480 GB of storage and a Mellanox InfiniBand HDR100 interconnect. Frontera delivers 23.5 Linpack petaflops, placing it fifth on the most recent Top500 list of the world’s most powerful publicly ranked supercomputers. Second was Longhorn, a newer, smaller system equipped with Nvidia Tesla V100 GPUs and IBM Power9 CPUs that delivers 2.3 Linpack petaflops, placing it 120th on the most recent Top500 list.
“In the last few weeks, we have gathered a lot of data and developed the most extensive set of simulations on SARS-CoV-1 and -2 spike proteins, made possible through our access to TACC’s facilities,” Moradi said in an interview with TACC’s John Holden. “It is one of the few supercomputing facilities in the world that allows for such a quick turnaround in terms of performing such large-scale biomolecular simulations.”
The researchers found that the two coronaviruses actually showed remarkably different behavior, despite visual similarities. COVID-19, for instance, bound to human cells much more tightly than the first SARS virus, with its spike proteins activating within microseconds – far faster than those on SARS. “Our simulations show clearly just how active these proteins become,” Moradi said.
This distinction between the two virus’ behavior had likely gone unnoticed due to measures taken to stabilize the virus for lab analysis, such as keeping it at very low temperatures. Moradi, however, used accurate temperatures and a more realistic environment.
“A synthesized lab structure is the equivalent of a still photo of two people,” Moradi explained. “It can tell you what something looks like, but it can’t tell you how it’s going to behave.”
Header image: the COVID-19 spike protein preparing to bind.