developing a single new vaccine or drug takes years to develop. Amidst the crisis, Artificial Intelligence (AI) is proving to accelerate this process. What role AI is exactly playing is explained ahead?
More than a hundred countries have been hit by COVID-19. The novel coronavirus is made up of genome sequences never detected so far. Hence, the world does not have a specific cure that could defeat the virus.
In an attempt to slow the spread, drug researchers across the globe are striving to develop a medicine. Some are focusing on creating vaccines that serve as the most powerful weapons for preventing communicable diseases. Others are working on developing drugs as a treatment for COVID-19.
The Interesting Role Of AI As The Drug Design Catalyst
In the past without AI, 30 scientists working in a lab for 30 years had the potential to develop only one drug. They had to go through a massive amount of data, follow trial and error methods, and come up with a solution. Today, the crunching of a huge amount of data is handled by AI. It is able to perform calculations to narrow down to important data at a much faster rate.
For instance, AI can help identify 50,000 molecules (from hundreds of thousands of other molecules) that researchers might be interested in. After these 50,000 molecules are identified, another AI software aids in further narrowing down this number to more specific ones.
In simple words, AI just takes a researcher to data he should be looking at – eliminating 98-99% of unwanted data. Thanks to the capability of AI. The development time that took decades can happen within a matter of months.
Furthermore, AI applications can also help researchers determine whether a particular molecule is safe to use. A range of cloud computing resources, supercomputers, and machine learning models are applied to guide researchers towards designing a vaccine. All of these systems are based on neural networks that are helpful especially in analyzing textual data.
Different initiatives tapping into AI to catalyze drug/vaccine design are explored next.
A Big Drug Development Activity In The UK
Seeing the urgency of a cure during this pandemic, pharma companies are trying to repurpose drugs. This means a drug developed to treat a particular disease may be effective in treating another. The human body comprises a complex biological structure. So, one drug may work for multiple diseases.
But repurposing drugs is a time-consuming process. Companies run trial and error experiments on already-approved drugs while having a basic understanding of the disease. In the case of COVID-19, the world is already using hydroxychloroquine and remdesivir to treat infected patients. The former is approved to treat Malaria while the latter is used for Ebola.
BenevolentAI, UK’s giant in the AI drug industry has amassed $292 million to create COVID-19 drugs with the help of AI. The company has experience in using AI to create medicines for the world’s deadliest diseases. This is their first attempt of using AI to find a drug for an infectious disease.
For some time now, it has been applying AI for repurposing existing drugs to defeat the novel coronavirus. As per sources, the company has already identified an already-approved drug to treat the virus.
In a month’s time, the company came down to six most promising molecules with baricitinib. Also known as a ‘JAK inhibitor,’ this medicine is originally used to treat rheumatoid arthritis.
Reasonably, there are many questions and concerns over baricitinib’s potential use as a COVID-19 treatment drug.
The National Institute of Health is soon going to test its effectiveness through its accelerated clinical trial. The drug is also being researched by other countries. Some of them include Canada, Italy, and the US among others.
Using Google’s AI
DeepMind, a UK-based AI company, and Google’s division is using its computing power to understand proteins of COVID-19. A virus is mainly made up of proteins. Identifying its proteins can help develop the treatment.
In any organism, proteins play a vital role. Understanding its three-dimensional shape could provide a blueprint of the virus causing the disease. As a result, it can potentially help in discovering a COVID vaccine.
The startup has already published its findings so that others can take benefit from it to develop a medicine. Recently, it offered data files consisting of the best guess of a handful of protein structures associated with the novel coronavirus disease.
DeepMind’s AI was capable of accomplishing this task with decades of effort by experts. This includes pioneering work and research by chemists, physicists, biologists, computer scientists, and data scientists. It dived into six decades of distinct science and data curation across the world – a global effort.
Digital Technology Supercluster
Ottawa in Canada assigned the nation’s five superclusters to connect with their 1800 members to find out ways to fight COVID-19. As a result, the Vancouver-based Digital Technology Supercluster committed to investing $60 million out of its $153 budget to develop partnerships. Since then, the supercluster has been reviewing over 300 submissions from its 500 members.
One of the partnerships it welcomed is between Variational AI and adMare BioInnovations Inc., an NGO involved in commercializing academic research. Besides, the supercluster has also given a green signal to three other projects being furthered in partnerships.
As far as Variational AI concerned, it is already using AI to accelerate the process of exploring an effective drug. The algorithm of the company is also capable of identifying approved drugs that could bind to COVID-19 proteins and defeat them. As the project involves identifying existing drugs, no clinical trials will be required. This will significantly cut the time.
On May 3, 2020, Canadian Prime Minister Justin Trudeau announced $175 million funding to Vancouver-based AbCellera Biologics Inc. This is in relation to the rapid development, production, and distribution of therapeutic antibodies. Clinical trials for the same will probably begin in the first week of July.
Prediction Of COVID Vaccine Development
Some existing AI software and platforms are already helping the drug industry. But some expert researchers are attempting to build brand new AI programs that can facilitate the development of the COVID vaccine. One such example of an AI platform is ‘Epitopes.world.’
A team of researchers from Harvard and Université de Montréal introduced this interactive platform. It is designed on an algorithm called ‘CAMAP’ that makes predictions for possible vaccine targets. Eventually, this allows researchers to identify the parts of the virus that are more likely to be exposed at the surface of infected cells (AKA epitopes).
Epitope.world will decrease the time and cost required in creating vaccine candidates. Project Lead Dr. Tariq Daouda who built Epitope.world with his team hopes. As per Pharmaceutical Research and Manufacturers of America, clinical trials take nearly six-seven years. And the cost of R&D reaches about $2.6 billion on average.
The algorithm CAMAP was originally applied to cancer immunotherapy. However, its propensity for learning immune system patterns proved it to be a perfect choice for uncovering virus vulnerabilities. Epitopes.world is not involved in synthesizing vaccines. Instead, it provides predictions that could help produce a list of epitope targets to examine.
For smart city leaders, in particular, this time is an opportunity to tap into the benefits of AI. As building resilient health infrastructure is a priority, AI’s role in drug development is the chance to prove its value.