How AI is fighting COVID-19: the companies using intelligent tech to find new drugs
The use of artificial intelligence (AI) in drug development has been hyped up for some time now – and the COVID-19 pandemic is finally a chance for these technologies to prove their worth. We highlight the companies that are using AI to develop treatments that could end the outbreak.
Pharma is facing a race against time to tackle the exponential rise of COVID-19 cases, and as a result there has been an unprecedented level of data sharing and cooperation in the industry to accelerate the process.
Drug repurposing is the most common strategy pharma companies are applying to address the urgency of the pandemic. Our body is a complex machine and sometimes one drug may work for multiple diseases. Perhaps the most successful example of repurposing is Pfizer’s Viagra in erectile dysfunction, originally intended to treat cardiac conditions.
Repurposing drugs with careful rational examination is a slow process. Companies are running trial and error experiments with already-approved drugs, having established safety profiles in humans, on the basis of basic understanding of the disease. When it comes to COVID-19, hydroxychloroquine (approved to treat Malaria) and remdesivir (for Ebola) are the two best-known examples of this so far.
In recent years, AI based drug-discovery has been much hyped as an effective way to speed up the drug discovery process. Traditionally, it takes 3-5 years to discover and optimise molecules with animal studies before they are tested in humans. Rational and directed repurposing efforts can take anywhere between 1-3 years. AI-based start-ups have long claimed to identify and design new drugs in a matter of days and months. This would save multiple years and billions of dollars to bring a new treatment to market.
Now seems like the perfect time for them to prove they can rise up to the challenge.
Many AI-drug discovery companies are working to speed-up a rational repurposing of the available drugs against the novel coronavirus.
In general, the success of AI platforms is dependent on the data that is used to ‘train’ the algorithms. Limited availability of data about 2019-nCoV, as the world is still studying the virus, can be a challenge.
The companies are surfing through hundreds of terabytes of data about other known viral diseases and drugs to model COVID and predict drugs. Some AI companies are also trying to design new molecules, in case no currently available drugs deem effective.
List of companies using AI for developing COVID-19 treatments
Repurposing existing drugs:
- BenevolentAI: The UK based company is known as a giant in the AI drug discovery industry. The company has been using AI to repurpose all existing approved drugs against the novel coronavirus. Within a month, they narrowed candidates down to the six most promising molecules with baricitinib, a drug rheumatoid arthritis drug, being most promising for treatment. They published their result in The Lancet, after which Eli Lilly got in touch and has now begun to clinical trials in US.
- Innoplexus: The Indo-German company began by evaluating the potential of therapies like Hydroxychloroquine and Remdesivir against 2019-nCoV by using data coming from patients. Their AI platform suggested higher efficacy for combination of chloroquine and tocilizumab (a drug for rheumatoid arthritis), chloroquine and remdesivir and a third combination of hydroxychloroquine with clarithromycin (an antibiotic) or plerixafor (antiretroviral). The company is validating all these combinations in vitro and in vivo. They are also generating novel candidates using their platform.
- Deargen: The Korean company, in collaboration with Dankook University, predicted the activity of available antiviral drugs against the novel coronavirus and the AI platform suggested atazanavir (a drug for HIV treatment) to have high potency. Abbvie is running a trial with its similar HIV-antivirals lopinavir and ritonavir in Wuhan. The company proposed a few more drugs which are under consideration.
- Gero: The Singaporean company identified 9 drugs using its AI platform. Among the top molecules predicted to have efficacy against the coronavirus are niclosamide and nitazoxanide, broad spectrum anti-parasitic and anti-viral drugs.
- Cyclica: The Canadian company announced screening of 6700 molecules that are FDA approved or at-least in Phase I human trials on their AI based drug repurposing platform MatchMaker. They have partnered with China’s Institute of Materia Medica for in vitro and in vivo assessment. The results are expected in May.
- Healx: The UK based company is using its platform to uncover bi-and tri-combinations of approved drugs against the virus. While the company’s capabilities lie in rare diseases, they are uniquely positioned to leverage the data on why mortality is higher with comorbidities of respiratory and cardiac systems. Candidates will be available in May for external testing.
- VantAI: The New York-based company is using a systems biology approach to understand the interplay between interactions of the viral proteins with possibly more than 500 human proteins over the course of infection. The company is aiming to identify drugs against targets that may have been overlooked by the other companies and has already uncovered several promising target pathways, for example blocking the virus’s activity in the Golgi apparatus (a viral re-packaging system in the human cell that helps in further spread), and is currently screening around 300 leads with a CRO to experimentally investigate impact on viral infection further. The results are expected by end of May.
Designing new drugs:
- Insilico medicine: The Hong-Kong based company ran operations and published 100 promising small molecules against the 2019-nCoV after a 4-day sprint. While their chemistry efforts were disrupted by lockdown, they have managed to synthesize two of the seven most promising molecules and plan to test them by early May.
- Exscientia: The UK based company partnered with the Diamond Light Source, UK’s synchrotron facility, to screen about 15,000 clinically ready molecules from Scripps Research Institute in California, US. It is also modeling data to understand how the virus works, with the first findings expected by early to mid-May. They believe that since their original repertoire is much bigger than the set of approximately 2000 approved drugs, it will help them lay the foundation for the next phase if current therapies fail to benefit.
- Iktos and SRI international: The French AI specialist Iktos plans to design novel molecules and has partnered with US-based SRI Biosciences to utilize their fully automated end-to-end synthetic chemistry system to synthesize and test the molecules.
Many in the industry believe that the use of AI is overhyped in drug discovery. If these AI discovered molecules prove to be successful, skeptics may have to bite the bullet. This is the chance for AI-based companies to mark the new decade by providing an efficient and agile discovery process. It is a chance to fulfill the visionary mindset of pharma companies to provide accelerated therapies at a price that every patient around the world would welcome.
About the author
Amandeep Singh is a life science consultant at MP Advisors.