AI-powered supercomputing has come to the fore in the European fight against COVID-19. The EU-funded, public and private consortium Exscalate4CoV (E4C) has escalated the fight against COVID-19 through supercomputer-driven drug discovery.
The E4C’s key aims are twofold – to develop an effective tool for countering future pandemics to be consolidated over time, and to identify molecules capable of targeting COVID-19. Its supercomputing platform generates and analyses 3D genomic models of COVID-19, and compares them to the structures of an array of molecules to identify possible treatments.
Having already tested 400,000 molecules, Raloxifene, a generic drug to treat osteoporosis – a condition that slowly weakens bones – has been identified as a possible COVID-19 treatment. Modelling through the supercomputer, and subsequent research, has indicated the drug could be effective in blocking the replication of the virus in cells, halting the progression of the disease.
Raloxifene usually prevents osteoporosis in postmenopausal women, but may also be used to reduce the risk of breast cancer. Its multifaceted physiological effects may not clearly pertain to COVID-19, and it is just one of billions of molecules that could be used to combat COVID-19. The capacity to test such drugs against COVID-19, and so quickly, has been made possible through AI-powered supercomputing.
Escalating computer power
The E4C platform has proved the keystone in European COVID-19 research efforts. Through X-ray analysis of the virus, the platform is able to model the different protein structures of the virus that are constantly evolving. New models received each week are translated into a digital structure to analyse against its repository of over 500 billion molecules.
E4C has transformed standard computing in healthcare, feeding on and supporting the data-driven turn in healthcare. Supercomputers run parallel, rather than serial, processing in a process known as ‘floating-point operations’ (‘flops’); computers can perform many operations at once, equivalent to having thousands of normal computers running in-sync.
The EU platform has 120 Petaflops of power; to perform the calculations the platform performs in one second, a human would have to perform one calculation every second for nearly four billion years. This has allowed it to already process 400,000 molecules against COVID-19, with five million more in the pipeline for the next phase of the project.
AI-powered discovery
The E4C project harnesses the ability of artificial intelligence to drive its supercomputing platform in its race to combat COVID-19. AI technology enables the E4C platform to effectively screen, process and analyse the molecules – it forms the brains of the platform, much as the supercomputing platform constitutes the body. The volumes of data produced by the billions of calculations performed by the platform requires new forms of processing power made possible through artificial intelligence.
New Horizons
Data-driven healthcare research and delivery has become a key focus for governments across the world to accelerate discoveries and realise transformational change. The tandem power and processing capabilities of supercomputers and AI could drive these new horizons in healthcare; the accelerated power of AI-based research may be critical in rapidly meeting new challenges posed by COVID-19.
The E4C platform is driving research through the EU’s research and innovation programme Horizon 2020, which brings together industry and public firms and researchers to combat the virus. The funding for the research consortium comes off the back of over €600 million of EU investment in high performance computing, much of which has been channelled into Horizon 2020. A further €50 million was pledged before the onset of the pandemic to develop a new network of AI research networks across Europe, promising new horizons for data-driven medical research.
