Supercomputing Power Towards Resolving COVID-19 Crisis

April 08, 2020

The world is at present witnessing one of the most difficult periods in its modern history. It is facing COVID-19 pandemic that has no medical solution in sight yet. Thousands of people are dying and many more are expected to succumb to the coronavirus in days and months to come. Obviously, all efforts are ongoing to save the world from any eventual catastrophe. Apart from undertaking passive measures like social distancing, which helps to break the chain of human transmission of the virus, scientists are trying their best to find an appropriate medicine/vaccine to fight COVID-19.

To find a medical solution to the COVID-19 crisis requires a multi-disciplinary approach. Actually,  any research and development in the fields of Biotechnology and Pharmacology is known to have always followed a multidisciplinary approach. Various sub-branches of science, for example, microbiology, genetics and biochemistry are also to be accommodated in such type of researches. In addition, the present-day research has significant amount of dependence on data. This is owing to the increase in technological capabilities for the ‘number crunching’ over the years. Development in computer science and data science is permitting researchers to derive maximum benefits for their research. Presently, researchers are found using supercomputers to assist them to find a medical solution to the COVID-19 crisis.

 Supercomputers have a long history. For many decades they are being used for various types of research. However, over the years a significant amount of change has taken place in the computing capabilities of the supercomputers. Essentially, this being a dynamic field for development, continuous up-gradation of speed for computing abilities and capability in handling the volume of data is taking place.

This is, however, not the first time the supercomputer is being used for such type of research. The Human Genome Project's sequencing effort was announced in 2004. It was IBM’s Blue Gene supercomputer which had played a major role in sequencing the human genome. This breakthrough realised major developments in the drugs sector and treatment methodologies. Blue Gene has helped to simulate approximately one per cent of a human cerebral cortex, containing 1.6 billion neurons with approximately 9 trillion connections, leading to a greater understanding of the human brain.

The supercomputing capability had been put in use for the assessment of the 2009 H1N1 flu pandemic (pH1N1). A very significant mathematical modelling effort was undertaken to understand the entire viral envelope of pH1N1. This was a major mesoscale simulation of pH1N1. It was one of the largest biological systems ever to be simulated on that level. The research community had taken the help of the Blue Waters supercomputer of the National Centre for Supercomputing Applications (NCSA) at the University of Illinois at Urbana-Champaign. This powerful simulation had helped the researchers to understand how flu proteins interacted with one another on the surface of the virus and also for undertaking statistical analysis of viral protein movement. However, this assessment was carried out after the pandemic phase was over. What is important at this point is to learn from these experiences and put the available computing power to quick use for finding a medicine and vaccine for coronavirus.

At present, some significant efforts are underway not only in the US but also in other parts of the world carried out by supercomputing agencies. In 2014, the US Department of Energy commissioned the ‘Summit’ for solving some of the world’s biggest problems using artificial intelligence (AI) and deep learning. Summit (OLCF-4) is the fastest supercomputer developed by IBM so far.  Scientists have already started using this machine for drug analysis for COVID-19. Simulations have been carried out on more than 8,000 compounds to screen for those which could be used as a new drug against the coronavirus. Some 77 compounds have been shortlisted for possible vaccine. Further research is ongoing.

Realising the importance of supercomputers to address the COVID-19 threat, around mid-March, the US White House declared the creation of the COVID-19 High-Performance Computing Consortium. There are three research institutions in the consortium -- the MIT, the University of California, San Diego, and Rensselaer Polytechnic Institute, New York. Numerous other agencies have joined in as well.  The purpose behind the formation of this grouping is to allow the research community involved in vaccine development an access to supercomputer facilities. So far, there are 16 supercomputers available in this network.

The supercomputer at the Joint Supercomputer Centre of the Russian Academy of Sciences (RAS) is also known to be working to develop drugs to fight COVID-19. In China, the country of the Coronavirus origin, various supercomputing facilities are put on the job to address the COVID-19 crisis. The Chinese are mainly using the supercomputer Tianhe-1 to predict the trend of the infection and also to evaluate the effect of prevention and control. More importantly, an AI-based interface is allowing the Tianhe-1 supercomputer to go through hundreds of images (chest scans of the people with the suspected infection) generated by computed tomography (CT) and produce diagnosis within 10 seconds. This has proved to be of great assistance to the medical professionals both in cities and also in areas where limited medical facilities are available. Interestingly, the researchers are also known to be conducting a large sample calculation experiment to provide accurate data support for decision-making on the resumption of public and private activities and other works in the country.

 The Digital Transformation Institute (DTI) is another supercomputer cooperative network made available for the use of scientists for research on COVID-19. This unit is a collaborative effort by Microsoft Corporation and various other US-based universities. This network has been established to take forward the new Science of Digital Transformation of Societal Systems. This agency also has a major focus on AI, Machine Learning (ML) and the Internet of Things (IoT). This group is run by two American scientists (of Indian origin) who have extensive experience working with the supercomputers.

In addition, the Texas Advanced Computing Centre (TACC) has been working on issues connected with various disease outbreaks for the last few years. This Centre has been working on the coronavirus outbreak too. Resources at this facility were used to predict the speed of the coronavirus outbreak when it was first labelled as an epidemic. In fact, TACC has an existing tool called the Texas Pandemic Flu Toolkit which has most likely been used to handle the COVID-19 crisis. This tool uses supercomputers to optimize emergency health care response. This system was developed during 2012 in response to the H1N1 swine flu crisis. The process involves simulating the spread of pandemic flu over an area of interest. This allows to forecast the possible number of hospitalizations and determines where and when to place health care resources to maximize saving lives. Now, an updated model is available to predict the outcome of COVID-19.

  A supercomputer could be a standalone system or a system of systems. The type of computing power available with one supercomputer could be gathered by joining together various smaller computers too. An agency called Folding@home is presently working to achieve such capabilities for undertaking research on disease control.  The word ‘Folding’ refers to the way human protein folds in the cells that make up the human body. Humans rely on proteins to keep them healthy and they assemble themselves by folding. But when they misfold, there can be serious consequences to a person’s health. The 'Folding@home' software allows people to share their unused computer power to this project. This is a distributed computing project for performing molecular dynamics simulations of protein dynamics. Its initial focus was on protein folding but has now shifted to biomedical problems like Alzheimer's disease, cancer and Ebola. Obviously, at present, the emphasis is more on COVID-19.

Currently, the inability of humans to medically address the COVID-19 infection is increasing the global crisis. At the same time, it is heartening to note that scientists are using all the available knowledge, resources and technologies to tackle this threat. It is important to note that various research teams all across the globe are studying COVID-19’s notorious “spike” protein. The purpose is to find a solution to deactivate the virus’ ability to infect humans and there can be seen a significant amount of dependence on the supercomputing capacities for this purpose. Timely investments made by countries like the US, Russia and China in the field of supercomputing and AI are proving highly beneficial at this juncture of COVID-19 crisis.

Note: Various online opensource materials were consulted for this article.  

Author Note
Dr. Ajey Lele, Senior Fellow, MP-IDSA, New Delhi. Views are personal.