As the numbers infected with COVID-19 continue to increase, health care systems around the globe do what they can to take care of the sick. IBM’s supercomputer, Summit, is one way that technology has been put to the test to study potential vaccines for the coronavirus.
The IBM AC922 Summit is the world’s most powerful supercomputer. The supercomputer is a part of the US Department of Energy’s (DOE’s) Oak Ridge National Laboratory (ORNL).
Because of its superpowers, the Summit has identified 77 small-molecule drug compounds that might warrant further study in the fight against the SARS-CoV-2 coronavirus. This is a positive step to understanding how to treat the COVID-19 disease outbreak.
The two researchers that man the Summit have performed simulations of more than 8,000 compounds. They have done so in order to screen for compounds that are most likely to bind to the main “spike” protein of the coronavirus.
If researches are successful, they will have gained amazing insight into how to stop COVID-19 from infecting host cells (that’s us!). The researchers have published their results on ChemRxiv.
Summit and the link to COVID-19
IBM’s Summit is housed at the University of Tennessee. It is in a room that is twice the size of a tennis court. The supercomputer is known for its studies of Mars landings and the origins of the universe. Now, it can add understanding the coronavirus to that list.
The US Department of Energy deployed IBM’s Summit computer to simulate how different variables react with the virus. That means running countless numbers of simulations and variables. This is something that computers are very good at doing, however, unlike any other computer, the Summit can run more simulations faster than other supercomputers.
The positive result of the Summit is that researchers have uncovered 77 compounds that could be used to help individuals be protected from COVID-19 infection. These compounds include potential therapies and medications. The compounds were found from 8,000 known potential compounds. A process that would have taken months in the past only took several weeks with the Summit.
How Super is the IBM Summit Supercomputer?
If you are wondering what “fast” results are from mega or supercomputers are, take the following comparison from Jeremy Smith, Governor’s Chair at the University of Tennessee who is the principal researcher in the study:
“Summit was needed to rapidly get the simulation results we needed. It took us a day or two, whereas it would have taken months on a normal computer.”
The Summit is capable of performing over 200 quadrillion calculations per second. That makes IBM’s Summit the “Formula One of supercomputers.” More importantly, it makes this supercomputer 1 million times faster than the most powerful laptop.
Many of these simulations require billions of calculations. This is a task that could take more than 500 years to complete on a single, powerful PC.
So, the difference is days v. months and years! That is how much faster the Summit can run simulations. This is very promising news indeed. But it is important to be patient with any kind of results such as these.
Smith further states:
“Our results don’t mean that we have found a cure or treatment for COVID-19. We are hopeful, though, that our computational findings will both inform future studies and provide a framework that experimentalists will use to further investigate these compounds. Only then will we know whether any of them exhibit the characteristics needed to mitigate this virus.”
Research Sharing and the Whitehouse
While the Summit is housed in Tennesse, Dario Gil, director of IBM Research, states that they have partnered with the White House. This partnership is an effort to share their findings with others. Sharing information with other research teams increases the chances of stopping the spread of the coronavirus pandemic.
Computers are invaluable when it comes to developing and testing predictive models. These models help to analyze and understand how the coronavirus progresses and is spread. Models also help to develop potential therapies and vaccines, as the Summit can run models that test the response to other similar viruses.
Gil adds to the earlier statement:
“These high-performance computing systems allow researchers to run very large numbers of calculations in epidemiology, bioinformatics, and molecular modeling. These experiments would take years to complete if worked by hand, or months if handled on slower, traditional computing platforms.”
The current partners of the new consortium include NASA, MIT, Rensselaer Polytechnic Institute, Lawrence Livermore National Lab, Argonne National Laboratory, Oak Ridge National Laboratory, Sandia and Los Alamos National Laboratories, and the National Science Foundation.
Why Do We Use Computational Models to Treat Illnesses?
Computational modeling is crucial for understanding the function of disease. The information gained from modeling can be used to improve how genetics and diseases behave, which makes it possible to better treat these things.
This emerging field is called “computational medicine.”
These models include:
- models of molecular networks
- physiological processes
- as well as modeling anatomic shapes layered with physiological function.
There is more than one way to execute the study of these compounds. However, a common thread is the use of quantitative models to understand altered structure and function in disease.
It is important to understand that modeling is only a first step in looking for answers. Once the models are created with computational medicine research, the model predictions may or may not be supported by results from the studies that take place afterward. However, once the predictions are tested, the models can then be revised based on the results. The continued refinement of results leads to a better understanding of the virus or disease.
Ultimately, the more studies that go on and are shared, the more data that supercomputers, like the Summit, can test. Collaborative research and supercomputing will lead to the most likely compounds that should be tested to treat COVID-19.
One of the keys is sharing results with others in the scientific community, something that world leaders are presently encouraging, given the state of COVID-19.
One such effort to share information and computing power is Folding@Home which aims at using shared computing to get results faster than the Summit.
The project was initiated by Stanford University to utilizes the spare CPU and GPU cycles of ordinary PC users around the world. Just like decentralized servers, the idea is to collectively run computer simulations related to the study of the coronavirus.
By combining efforts, individual computers have been able to support the research as well.
“To put that in perspective, that’s more than 2x the peak performance of the Summit super computer!”
says Bowman in a recent Twitter post.
The computations run are to understand how the atoms move with a protein. In order to treat COVID-19, they must be able to find a protein that will bind with the COVID “spike”. The ultimate goal is to discover “druggable pockets.” Druggable pockets may lead to potential cures to COVID-19 or similar diseases. Nothing is certain, but many discoveries are made in the search for something else. No research in the field is wasted effort!
Viruses and Proteins
To understand COVID-19, studies of other similar viruses are used as research starting place. For example, the Ebola virus. One study that simulated a protein from the Ebola virus. With Ebola, they found a site (or a protein) in the virus that could be treated with drugs.
This led to the research team’s ability to perform experiments that confirmed the computational prediction. Which means they are now searching for drugs that bind this newly discovered binding site.
Understanding the proteins of a virus is crucial for understanding how it is able to suppress an immune system, which then allows the virus to reproduce. Proteins are molecular machines that perform many functions we associate with life. And although viruses cannot survive without a host, they are made up of proteins, just like plants and animals.
Proteins sense the environment and play a central role in all functional structures. To understand how something works, a scientist must understand the protein make-up. The way that a protein’s components are arranged determines the way the virus functions. Therefore, understanding a new virus, such as COVID-19 is essential to learn how to treat it.
Although there are many experiments that can determine the protein structure of a virus, this does not mean the entire protein structure is revealed and understood. Instead, experiments often only lead to a “snapshot” of the protein’s typical shape. With many ways for the protein to behave, it is important to see what some of these behaviors are. Unfortunately, it is only the known structures of a protein that can be studied.
Chinese researchers were able to learn that COVID-19 infects the body in a similar way as SARS (Severe Acute Respiratory Syndrome). The SARS virus managed to spread to 26 countries during the SARS epidemic in 2003. It is the similarities between coronavirus and SARS that have facilitated much of the research that has gone on so far.
Viruses and Computations
As mentioned earlier, because of the quick work of Summit, researchers now have 77 small-molecule compounds to begin the next stage of research with. These 77 compounds came from 8000 known compounds that react with a SARS-like virus.
The 77 compounds are things such as medications and natural compounds, that they suspect may valuable, and will now be used for experimental testing. Based on the simulations obtained from computational models, they now understand that compounds bind to regions of the “spike”. Knowing what the “spike” is essential for understanding how the virus enters the human cell.
Science recently published a study that revealed a highly accurate S-protein model. Using findings from the study, the next research team will be able to rapidly run more computational studies; but with the new version of the S-protein. The new results may change the ranking of the chemicals researchers currently believe are the most useful.
Narrowing the compounds using computational models is only another step in the research process, however. Researchers emphasize the need to test the 77 compounds experimentally. This next phase is essential before any determinations will be made about how successful these compounds really are.
Final thoughts on supercomputing and the fight against Coronavirus
The Summit and projects like Folding@Home are amazing examples of the efficiency and capacity offered by supercomputing. They represent a massive initiative in the global fight against COVID-19. The 77 compounds found so far are a bright spot in the research process for treating Coronavirus. However, it is important to keep perspective on the outcome of this positive research.
As scientist Jeremy Smith points out, these computations must be followed by experiment. Computational screening essentially “shines the light” on promising candidates for experimental studies. Narrowing the field is essential to verify certain chemicals that may combat the virus.
With the help of IBM Supercomputer Summit, these results can be narrowed. What once took months and even years, now only takes weeks!