The Role of AI in Drug Discovery
By Veni Dole (SEAS’25, Analyst 2022–23)
Drug discovery involves manipulating genes, molecules, and cells to then affect the human body or other biological systems in a desirable manner. To understand the role of artificial intelligence, AI, in genomics-based science, it must be thought of as code: in which genes can be manipulated analogous to lines of code to build a program. Therefore, just like AI can be used to find patterns in code, it can be used to find patterns in genes.
When looking at the genes relevant to the target of a drug, there is a large amount of data to process in that there are infinite ways that genes and molecules can interact and thus infinite possibilities for how to manipulate them. Traditionally, scientists must make decisions on how to manipulate genes based on prior research and long experimental processes with much trial and error. However, AI can now provide scientists with smart options at various stages of the drug development process. For example, AI can be used to identify genomic targets, map previous research, predict protein structures, and flag potential risks in a technology or a patient to whom the drug is being administered.
The benefits of applying AI to drug discovery can be viewed quantitatively: by traditional methods, developing a new drug can cost from $1–2.5 billion. However, according to research by Morgan Stanley, the application of artificial intelligence could create a more than $50 billion opportunity through the creation of 50 additional drugs over 10 years. The time needed to develop a new drug could be reduced from 4–5 years to less than 10 months. Furthermore, the technology could also yield a 20% to 40% reduction in costs for preclinical development. As an industry expanding at an annual rate of 40%, headwinds include a post-COVID-19 increase in interest in and funding of biotechnology research and drug development. However, a negative public stigma surrounding artificial intelligence and genetic engineering remains. Additionally, even with the integration of artificial intelligence into drug development, issues of the process from a business standpoint remain relevant, including a long cycle, the need for clinical trials, drug regulations, and the impossibility of fully modeling the human body. The vertical also has a high barrier to entry in the cost of developing AI tailored specifically to a biological company’s needs or adopting existing AI for new purposes.
Image source: https://www.bcg.com/publications/2022/adopting-ai-in-pharmaceutical-discovery
Overall, AI can improve the efficacy and efficiency of the drug development process. As a result, the time required for and subsequently the cost of the process significantly reduces. In combination, these factors allow for researchers to address more diseases than previously possible. AI also opens the door for improved personalized medicine as it can process both the molecules of a drug and the genes of a patient to make connections between the two. Business models for the vertical include collaborations by companies developing relevant AI with pharmaceutical companies as well as in-house drug development. Major players currently include Benevolent AI, Recursion Pharmaceuticals, Schrodinger, and Kymera Therapeutics, among others. Companies such as AstraZeneca, Johnson & Johnson, Pfizer, Microsoft, IBM, and Google are also collaborating with groups developing biotechnology-related AI.
Sources:
