High-performance computing is the workhorse driving the understanding of COVID19 today. It helps universities and governments analyze the vast amount of data in a short time to understand and curb the current outbreak.
The researchers have linked the origin of this virus to a seafood market in Wuhan, China. However, the outbreak in New York appears to have European roots. This has also fueled the COVID19 outbreaks across the nations, including Arizona, Louisiana, and California. These links were identified by sequencing the genome of SARS-CoV-2 to track mutations. Understanding this mutation is essential for developing a vaccine. But, this research demands tremendous computing power, since the average genomics file is several hundred gigabytes.
Researchers have identified 69 promising sites on the proteins around the COVID19 that could most probably turn out to be drug targets. One of the famous supercomputers, Frontera, is working to build an all-atom model of the COVID19’s exterior element, which encircles about 200 million atoms, which lets simulations happen around successful treatment.
Besides, some are building 3D models of the virus’s proteins only to identify specific places on the surface that could be affected by drugs. They leverage the power of molecular docking, which is reinforced by high-performance computing to foresee the connections between molecules and proteins.
A cryo-electron microscope must consider taking several thousands of molecular images to model this protein. Without using any of the high-performance computing systems, one might spend several years turning those images into a model and simulating the drug exchanges.
The Summit supercomputer, which is known to finish about 200,000 trillion calculations per second, has already screened beyond 8,000 chemical compounds to see their attachment to the spike protein, classifying 77 that might fight against the virus.
The usage of high-performance computing to understand the effects of COVID19 is more than the genetic or molecular level. For instance, we know that neural networks are trained today to identify signs of the virus in chest X-rays. Large-scale AI and high-performance computing can be utilized on the same system to input a considerable amount of data to the AI algorithm and make it much smarter.
The possibilities of high-performance computing are nearly endless. One can create the fluid dynamics of a forcefully exhaled group of particles, based on the size, speed, and spread. Similarly, one can build the system to identify the spread of viruses through air ducts and ventilation systems, especially in assisted living spaces and nursing homes with vulnerable populaces.
Besides, the researchers can simulate the supply chain of any given product, and its impact if any supplier is detached from the calculation or the spread of the disease based on multiple levels of social distancing.
We know that the current COVID19 crisis is complex and rapidly evolving. Therefore, gaining a robust grasp of the situation requires the ability to collect data on the novel virus and run various models and simulations. One can achieve this only with sophisticated and distributed compute capabilities.