Despite the COVID-19 pandemic being a colossal tragedy to humankind, there were also momentous scientific breakthroughs peppered throughout that period, especially in the field of artificial intelligence (AI) in biomedical science.
The first breakthrough was with Google’s protein prediction server called AlphaFold. A throwback on basic biology, proteins are a long chain of atoms that could fold into various three-dimensional shapes. The precise way of how these atoms fold would determine its function as enzymes, antibodies or other components necessary for human life. Knowing these shapes are important for understanding human biology, how diseases happen and for drug development.
However, determining the structure of these proteins requires several pain-staking steps, collectively known as X-ray crystallography. Conventionally, scientists have to re-program bacteria or cells to produce their proteins of interest and make crystals out of them. These crystals would then be bombarded with X-ray. X-ray crystallography is as much of a science as well as an art. To-date, there are roughly 100,000 protein structures determined using this method. In spite of that, this number pales in comparison to the billions of known protein sequences. This is where AI truly shines.
By analyzing the multitude of protein structures and identifying patterns of how protein folds, AlphaFold could very accurately predict various proteins. Recently, Janies’ Lab from University of North Caroline had harnessed the power of AlphaFold to study whether the Omicron variant of SARS-CoV-2 could escape neutralizing antibodies. Such computational tools are important for scientists to quickly study diseases such as COVID-19 and provide medical solutions!
The second breakthrough came with the UK-based drug discovery company called BenevolentAI. BenevolentAI revolutionizes drug discovery by using AI to sieve through numerous chemical compounds in their databank to shortlist likely drug candidates. For example, early on in February 2020, scientists in BenevolentAI predicted that a rheumatoid arthritis drug called Baricitinib could be used to treat COVID-19. In April, clinical trials (the COV-BARRIER trial) was already well underway and showed that Baricitinib could indeed reduce mortality of hospitalized COVID-19 patients by 38%. As of today, Baricitinib is deemed as one of the COVID-19 drugs approved by the World Health Organization. Currently, one of the global drug discovery powerhouse AstraZeneca is in partnership with BenevolentAI to discover drugs for lupus, lung disease, kidney disease and heart failure.
AlphaFold and BenevolentAI are just two of the many example showcasing how AI has changed the landscape of biomedical science and biotechnology. University academicians and undergraduates in biology-related fields should constantly keep abreast with these technological developments so that we do not get left behind. Gone are the days where biologists are only expected to don on lab coats and work in labs. Sometimes, with the right knowledge and right tools in hand, one can change the world by sitting in front of a computer.
Dr Xavier Chee is a lecturer at Swinburne University of Technology Sarawak Campus. He is contactable at xchee@swinburne.edu.my