The Intersection of Chemistry, Robotics, and AI in Drug Discovery (BIO 2024)

Published on: 

Lee Cronin, founder and CEO of Chemify, discusses the combination of digital chemistry, robotics, and AI in the drug discovery space.


As BIO 2024 moves into full swing, Lee Cronin, founder and CEO of Chemify, a UK-based company specializing in building up digital chemistry for chemical research, discovery, and manufacturing, discusses how the combination of digital chemistry, robotics, and artificial intelligence (AI) are advancing the drug discovery space.

According to Cronin, the concept behind digital chemistry is to connect chemical reactions to synthesize molecules in standardized hardware, or in a standardized way. This involves “using a programming language so that we can very faithfully reproduce chemistry, on demand and also to a given quality,” Cronin states. The process of computation in digital chemistry is akin to running code on a computer, he emphasizes. “The concept of digital chemistry … is pioneering this ability to turn code into molecules reliably at scale, on demand, and also for the creation of quite complex molecules, and even APIs,” he adds.

The combination of programming language, robotic architecture, and the use of data and AI is creating a new approach to not just making molecules but discovering and doing process optimization, practically in one tool chain, Cronin explains. “What this means is that we should be able to go from the discovery process to dreaming up new molecules to then the optimization and scaling,” he says.

It is not necessarily easy to dream up a molecule in-silico and then translate that into producing that molecule on demand because some aspects may not be worked out, or the yields might be low, or there may be other blocks, Cronin points out. “What Chemify has done is we've built a series of approaches that allow us to digitize chemical reactions and have them ready on demand—a bit like building blocks,” he states. This approach allows the user to pull a code and go from a chemical reaction A to B or from a chemical reaction A to D, depending on preference. “[This approach] allows you to go from your target that you dream of, to then work out what particular reactions you must do ahead of time, and, more importantly, allows you to restrict your search of chemical space when you're [early in the research process], when you're dreaming up new molecules to reactions you know how to do, [and] so this increases the speed that you can discover new molecules and make new molecules, and go through the cycle of design–make–test–analyze,” Cronin explains.

Of particular significance is that once a new molecule has been made, once the code has been established, that code can be reused and optimized to scale further down the line, Cronin points out. “You've got this programming language and these reactions and the robots that run it; [therefore], you're able to update your process continuously with the weight of the insights that you get,” he adds. With these chemical reaction insights, the data and methods can then be processed through AI, allowing for optimized yields, Cronin states.

View the video above for the full interview with Cronin. Cronin is speaking at the New Tools, New Times: Chemical and Structural AI for Drug Discovery panel at BIO 2024 in San Diego, Calif., which is being held June 3–6.