Whilst not a new application, the advanced adoption of technology to power computer simulations of protein folding is helping scientists combat COVID-19.
Why simulate protein folding?
Put simply, protein folding is the process in which a molecular structure assumes a functional 3D shape. Understanding the 3D molecular structure of a virus can be crucial information for scientists as with it, scientists in theory have a better opportunity to limit the spread of the disease.
In other words, trying to understand the virus without a 3D molecular structure is like trying to build a house without the blueprint design.
The difficulty is that the 3D molecular structure of a disease can be very difficult to predict because of the near unlimited shapes this can take. There are three main denominators that scientists have to account for: The folding code (sequence of amino acids), structure prediction and folding speed. This difficulty is known as ‘the protein folding problem’.
Therefore, running computer simulations of protein folding can become crucial for scientists as it can test different structures far quicker than human trial and error. Without computing simulations, testing all of the structural possibilities of a molecule would take longer than the current age of the universe – this is known as Levinthal's paradox.
Even with computer simulation, there are still countless possibilities that need to be tested in order to find an accurate model of the virus. These tests are a ‘sketch’ that could show a ‘way in’, but scientists must continue to examine further any possibilities that arise from the software.
This article also summarizes the protein folding problem in relation to computing simulation if you would like more information.
Folding@Home: A worldwide supercomputer
Folding@Home at Stanford University allows computers to donate their ‘idle’ GPU to run computer simulations. When the simulations reveal a ‘way in’ to the protein, researchers can qualify this and use in the combat against COVID-19. It streamlines simulations for scientists, and it works.
Their research has made tangible steps for the science community, contributing to over 100 research papers. In one example, their simulations uncovered an alternative structure for the Ebola virus that was previously considered ‘undruggable’ and they have continued to champion long-term work for Alzheimer’s and Parkinson’s.
Their partners include Intel, AMD and NVIDIA, whose components support Folding@Home. NVIDIA recently posted to Twitter encouraging their gaming customers to test their GPU power with the software.
Theoretically, the current estimated computer power available to run protein folding simulations is ten times faster than IBM Summit, the world’s fastest super computer.
The key for Folding@Home is the example it sets for participatory science. It wholly depends on individual users and companies to drive research and a cross-border commitment to understanding the virus more clearly.
Users include Brazil-based oil company Petrobras, Google, YouTube communities, ECGA amongst other computer hardware manufacturers, and individual users who are all contributing to this computer power.
In the United Kingdom, the Hartree Centre, a high performance computing, data analytics and AI research facility, are also running the front end of the COVID-19 target finding.
As the UK remains in lockdown, it can be stressful to know how to help fight combat COVID-19. As the Hartree Centre showcases, Folding@Home offers users the opportunity to contribute towards fighting COVID-19 and give back to the science community who are actively fighting the disease.
The emergence of AI
AI takes protein folding one step further.
DeepMind have released structure predictions for several under-studied proteins associated with SARS-CoV-2, the virus that causes COVID-19, which have been generated using the latest version of their deep learning system, AlphaFold. They did this through training neural networks to predict the shape of a protein from its genetic sequence. These neural networks would be optimized through ‘learning’, where small incremental improvements would be made with each simulation.
DeepMind stress that that while their structure prediction system is still in development and they cannot be certain on the accuracy of the structures it has provided at this stage, they hope they may add to researchers’ understanding of SARS-CoV-2.
However, this application indicates how AI could be utilized by the scientific community in the future.
This would not be possible without collaboration and innovation shared between the science and technology communities seen since the outbreak of COVID-19. It is heartening to see the tangible differences being made to counter the outbreak through this.
Importantly, the research mentioned here is underpinned by open source data which promotes research. Much of the protein folding has relied on the genomic blueprint released by China early in the pandemic. Folding@Home makes their work open to the scientific community, whilst DeepMind’s release has been posted with open access, continuing a culture of collaboration as default within the science community.