Among all possible chemical compounds, it is estimated that between $10^{20}$ and $10^{60}$ may hold potential as small-molecule drugs. Evaluating each of these compounds experimentally would be far too time-consuming for chemists. Thus, in recent years, researchers have begun using artificial intelligence to help identify compounds that could make good drug candidates. One of those researchers is MIT Associate Professor Connor Coley, who develops and deploys computational models to analyze vast numbers of possible chemical compounds, design new compounds, and predict reaction pathways that could generate those compounds. "It’s a very general approach that could be applied to any application of organic molecules, but the primary application that we think about is small-molecule drug discovery," he says.
Coley's interest in science runs in his family, and he graduated high school at the age of 16. He chose chemical engineering at Caltech to combine his interests in science and math. During his undergraduate years, he also pursued an interest in computer science and worked in a structural biology lab using Fortran to solve protein crystal structures. After graduating, he came to MIT in 2014 to start a PhD, focusing on optimizing automated chemical reactions by combining machine learning and cheminformatics.
Coley's work included a DARPA-funded program called Make-It, aimed at using machine learning and data science to improve the synthesis of medicines from simple building blocks. He accepted a faculty position at MIT at age 25, drawn by the resources and interdisciplinary support for AI and science.
After a postdoc at the Broad Institute, where he worked on identifying small molecules from DNA-encoded libraries, he returned to MIT to build his lab. His lab aims to deploy AI to synthesize existing compounds and design new molecules. They have developed models like ShEPhERD, which evaluates potential new drug molecules based on their three-dimensional shapes, and FlowER, a generative AI model predicting reaction products based on fundamental physical principles.
Coley emphasizes the importance of grounding machine-learning models in an understanding of reaction mechanisms, akin to expert chemists. His students are also involved in optimizing chemical reactions through various research threads.
Blogger's Review: Professor Coley's research exemplifies the profound integration of AI in the chemistry domain, particularly in drug discovery. By combining machine learning with chemical reaction mechanisms, the efficiency of new drug development can be significantly enhanced, paving the way for innovative pathways in the pharmaceutical industry. This intersection of fields is worth watching, as it may reshape the landscape of drug development.