IBM's HPC Consortium And The Impressive Anti-COVID-19 Effort


Earlier this week, I was interviewed on the use of supercomputers to fight the COVID-19 outbreak, and I thought it might be interesting to talk about the progress so far. 

One of the largest efforts in terms of resources, companies, and results currently being undertaken to fight the COVID-19 outbreak is IBM’s HPC Consortium.  It is arguably one of the largest efforts ever undertaken by the industry since Y2K to address a single exposure. Starting over two months ago and now consisting of more than 56 teams, including those from other large technology firms like Microsoft, Google, Amazon, HPE, and including the national laps at NASA, NSF, and several Universities, this is an incredibly impressive effort. 

This effort was also not done with any large contracts. It was a concerted effort across the industry that resulted in an analysis that would have taken lifetime had it been done by human hand and not with High-Performance Computers (and many of the top supercomputers). 

This effort was recently enhanced by the Partnership for Advanced Computing in Europe (PRACE). This group brought to the fight additional resources. Those resources included the Swiss National Supercomputing Centre’s Piz Daint, the 6th ranked supercomputer in the word, and the UK Research and Innovation’s supercomputers including Archer a 2.55 Petaflop monster based at the University of Edinburgh. As as well the Technology Facilities Council’s and Biotechnology and Biological Sciences Research Council’s HPC resources.  

Using these resources, these teams have made impressive progress, but we don’t yet see much of that progress because, often, human trials and testing (which happen at human speed) has to happen first. But let’s talk about the impressive work this Consortium has already done.

A Potential Atomic Covid-19 Force Field has been identified. Yes, like you, I initially thought someone was having me on. That is not the case.  Researchers Thomas Cheatham and Rodrigo Galindo from the University of Utah had been already studying how the potential energy generated by atoms give off charged “force field” that can attract or repel molecules using AMBER (which is a molecular simulator). 

Using the IBM Longhorn (Power Based) supercomputer, these researchers generated 2,000 models of compounds that could affect COVID-19 ranked by force field power. The result now will be used to help design better peptide inhibitors of the main protein-copping enzyme to stop a COVID-19 infection.  (I don’t know about you, but they had me at “force-field.”)

Natural Plant-Based anti-virals have been identified. India had an extremely rich data resource of around 3,000 medicinal plants, and anti-viral plant extracts already in use for medicinal purposes.  Novel Techsciences, an Indian company, is using the supercomputing resources to identify phytochemicals from these plants that could act as anti-viral drugs against the class of SARS-COV 2 protein targets. This research is being done to develop a prophylactic treatment (basically a COVID-19 preventative) against the virus that will overcome the anticipated future COVID-19 drug-resistant mutation.  (Think of this as a response to the anticipated 4th or later wave of infections). 

COVID-19 accurate indoor spread analysis is being done by the researchers at Utah State University in collaboration with Lawrence Livermore National Laboratory using Consortium Provided Supercomputers.  The result will help the Consortium better understand how virus-laden droplet clouds are transported and mixed within indoor environments like hospitals by air turbulence.  This effort, in turn, should help resolve the current concerns around the air-born transmission of this virus. 

New drug compounds that can mitigate or kill the virus are being researched by Innoplexus in Germany. They are using an HPC class AI to accelerate drug discovery. Part of this work is the design of brand-new molecules that could lead to new therapeutic treatments.  And they have discovered five at the time of this writing, that looks promising.  Innoplexus’ CTO Stratos Davlos and his team are doing impressive work.

The identification of genetic traits that could make a person more susceptible to COVID-19 to allow for them or their caregivers to protect them better is being undertaken by NASA.  Bio-scientists Viktor Stolc and David Loftus are trying to define risk groups for COVID-19 through genome analysis and enhanced DNA sequencing to develop accurate risk modes to initially find patients best suited for clinical trials of vaccines and anti-virals. 

Wrapping Up:  Decades Of Progress In Weeks

While this all makes me wonder what this same level of effort would do to the more traditional annual flu epidemic, or cancer for that matter, this assault on COVID-19 is unprecedented in terms of computing power and technology resources. And while many of the results won’t impact us until human trials are completed, we are learning a lot that will make it increasingly unlikely the world will ever have to go through another global medical catastrophe like this. 

And, I expect, before we are done, we’ll have a basis to understand better many of the other illnesses that shorten our lives and deprive us of our loved ones. I also expect we’ll eventually come up with a way to simulate the trials that are slowing this process down accurately, and that will be a huge game-changer. 

So if you are wondering what IBM and some of the other most powerful technology companies in the world are doing during this crisis, wonder no longer, they are having a massive impact, you just won’t see much of it until the slower human trials are complete.