As the COVID-19 pandemic continues to wage global disruption, the race is on to study the virus behind the pandemic: SARS-CoV-2 and find an effective drug or vaccine.
One of the technologies that is being most utilised is artificial intelligence (AI). AI is defined as the ability of a computer to rapidly think and learn, utilising machine learning to analyse large amounts of data. AI then can model predictions, screen virtually and develop insights that can impact R&D and assess patients.
AI is being used by many companies in a variety of ways to help better understand the virus and find novel solutions against it, including the identification and screening of existing drugs that could be repurposed, aid clinical trials, manipulate trial data and scour through electronic medical records (EMRs). AI has the power to sift through vast amounts of information and come out with rapid actionable results.
One of the most promising applications is the repurposing of existing drugs that could be effective against COVID-19. AI methods can be used to probe three-dimensional pockets of critical proteins in the viral infection pathway and analyse existing drugs that could fit into that pocket in order to generate hypotheses of whether a drug may be effective against the virus.
Another way AI can help scientists understand the COVID-19 pandemic is to scour through the electronic medical records of patients to determine if there are patterns of medications or pre-existing diseases they have that can help predict their disease prognosis. Russ Altman, co-director of the Stanford Institute for Human-Centered AI held an information session in April to answer questions about AI and COVID-19. He explained that if a molecular prediction points to a small-molecule drug and medical records can confirm that patients on that drug for another condition are experiencing lesser morbidity to the virus, that should be investigated. It would suggest that you are on the right track and that the drug might be useful.
An example of how AI has proven useful in the fight against COVID-19 was when BenevolentAI searched through existing drugs and identified Olumiant (baricitinib), Eli Lilly’s rheumatoid arthritis drug, as a good candidate. As a result of this, BenevolentAI and Eli Lilly announced that the National Institute of Allergy and Infectious Diseases had started a randomised-controlled trial to examine whether the drug can stop SARS-CoV-2 from infecting the lungs and reduce inflammatory damage.
BenevolentAI started researching COVID-19 in January following the outbreak in Wuhan and the team used medical and scientific literature to provide clues on how biological processes affect the virus. The team, led by Peter Richardson, found 370 known compounds that interact with the kinases that mediate the way in which viruses enter the cells, and this was eventually whittled down to baricitinib due to affinity, patient tolerance and known drug-drug interactions.
Innoplexus is another company using AI to investigate drug repurposing, who have identified that the combination of an anti-malarial drug (hydroxychloroquine) with remdesivir, or a rheumatoid arthritis drug (toclizumab) could be effective against COVID-19.
Novel drug design
Another way that companies are using AI is to identify new drug candidates. An example of this was announced in a recent collaboration agreement between Iktos, a French AI company and SRI International, a not-for-profit research institute to develop novel antivirals. SRI uses automated systems to design drug molecules using AI to identify the best synthetic route. SRI started research into targeting endonucleases in influenza about two years ago, which produced a compound and data set that could be used to build an AI model for COVID-19.
The benefits of AI are not limited to drug design or repurposing however, other companies are using these technologies to provide tools that can aid in diagnosis and contact tracing. This includes the work of Qure.ai who are using their tools to screen chest X-rays for signs of the disease and developed a platform which can be used for contact registration and tracing, and mapping COVID-19 hotspots.
AI has impacted the way clinical trials are being conducted as many have had to be digitalised due to social distancing. A company called Medidata demonstrated the use of synthetic control arms (SCA) which can create arms for studies based on historical control data matched to baseline characteristics of an experimental arm. A SCA can help with studies that have trouble enrolling participants and many patients in the control arm of a trial drop out.
While AI is a useful tool in the fight against COVID-19, Altman warned that it is just a piece of the technology puzzle, not a miracle worker. AI works in combination with other technologies in contributing to a solution that would have taken longer to come otherwise.