Drug discovery in the Era of Artificial Intelligence
From Alexa to Siri, facial recognition to voice commands, artificial intelligence – AI is gradually becoming an essential part of our life. But the concept is not new, back in 1950s famous British scientist Alan Turing suggested why can’t machines think like humans and make decisions on their own and this was perhaps the birth of the concept of AI (1).
Since then, technology has evolved exponentially, and AI is being employed in many different fields such as autonomous driving and manufacturing robots. One of the areas that are rapidly growing in competition is the pharmaceutical industry. Big pharma companies already collaborating or buying tech companies developing AI. Pfizer is exploiting IBM Watson, a program that uses machine learning to boost its hunt for immuno-oncology drugs. Sanofi is using an AI platform developed by UK start-up Exscientia to identify potential drugs for metabolic-disease (2).
Many of the drugs are small-molecule chemicals, and finding that correct molecule is a lengthy process that costs around $2.8 billion on average. These small molecules often bind to proteins through specific configurations and bonds. Identifying that exact molecule with perfect configuration and bonding property from a pool of millions of compounds is something AI can do much faster and better than humans. That’s why it can reduce the cost of development by as much as up to 70% (3).
So where is the future taking us? Well, we might see more and more pharma companies power their R&D with AI. This means the drug development process will be faster and faster – with strong competition, resulting in even more drugs available for patients. Perhaps one day, we might see a fully autonomous robot in the lab doing all the research that humans are doing now. Work for this has already started and one good example is the Adam, a robot scientist who performs experiments on yeast and analyzes the result (4). So, stay tuned for the next level of drug discovery!
Written by Zoljargal Baatarkhuu
(1) – https://sitn.hms.harvard.edu/flash/2017/history-artificial-intelligence/
(2) – https://www.nature.com/articles/d41586-018-05267-x
(3) – https://www.businessinsider.com/ai-machine-learning-in-drug-discovery-development-2020?IR=T
(4) – https://science.sciencemag.org/content/324/5923/85/tab-figures-data