Published March 10, 2026
Do you have any updates on the story? Please submit updates here. ›
Researchers have developed a machine learning-based system that can quickly predict the outcome of complex chemical reactions used in drug discovery. This has the potential to save time and money in the drug discovery process. This tool focuses on asymmetric cross-coupling reactions, which are important for generating the correct “handedness” of drug molecules, and can make accurate predictions based on limited training data.
why is it important
Drug discovery is an incredibly time-consuming and expensive process, often requiring extensive trial-and-error experiments to find the right molecular structure. This new AI-powered tool has the potential to greatly streamline this process and reduce the amount of physical laboratory work required by allowing chemists to quickly screen thousands of potential reactions and identify the most promising path forward.
detail
The researchers trained a machine learning model based on data from just a few published studies of asymmetric cross-coupling reactions used to construct complex drug molecules. The model was then able to accurately predict the outcomes of hypothetical reactions involving a variety of catalysts, ligands, and substrates, even when they differed significantly from the original training data. This allows chemists to quickly explore broader chemical areas and identify the most promising routes for further research, potentially saving weeks or months of laboratory work.
- The study was published in accelerated preview in Nature on February 11, 2026.
players
Simone Galarati
Co-first author of the study and joint postdoctoral fellow at the University of Utah and the University of California, Los Angeles.
Erin Butch
co-first author of the study and a doctoral student at the University of California, Los Angeles.
Matthew Sigman
chemist at the University of Utah and co-author of the study.
Abigail Doyle
is a chemist at the University of California, Los Angeles, and co-author of the study.
Do you have any photos? Submit your photo here. ›
what they are saying
“To understand new reactions, we sometimes use sophisticated physics-based computational chemistry tools. However, these tools are too expensive to predict thousands of potential new molecules. We wanted to train a statistical model that was ‘smart’ enough to accurately predict untested reactions, but was as cheap as possible.” ”
— Simone Gallarati, co-lead author (mirage news)
“Most AI requires huge amounts of data to train a model. This is a problem in chemistry where getting high-quality, large datasets from experimental work is very expensive and very time-consuming. The best part of this tool is that you can collect a small amount of data, build a reasonably good model, and accurately predict reactions that are known. You can also transfer the predictions to reactions that the model hasn’t seen yet.”
— Matthew Sigman, Chemist (mirage news)
“As a lab-based chemist, this tool is invaluable in saving me the time I spend running experiments. For example, instead of running 50-60 reactions, I can now run them 5-10 times, potentially saving weeks or months. Each reaction component I test in the lab must be purchased or made from scratch. This tool significantly reduces the amount of money I would normally spend on materials.”
— Erin Bucci, co-lead author (mirage news)
what’s next
Researchers plan to continue improving and extending the capabilities of machine learning models to make them even more powerful drug discovery tools.
Take-out
This new AI-powered system represents a major advance in the field of computational chemistry and demonstrates how machine learning can be leveraged to dramatically streamline the drug discovery process and reduce the time and cost of developing new drugs.
