A new machine-learning technique can predict how molecules will react with one another under varying conditions.
The method could allow the researcher to extract basic chemistry insights and speed the development of commercially important molecules, like pharmaceuticals.
“We often find surprises in datasets, and that becomes the launchpad for understanding,” says Alpha Lee, a faculty member at Cambridge Univ. and the chief scientific officer of PostEra, a medicinal chemistry company that he co-founded.
Lee and his colleagues call their method the “reactome,” playing off ideas from genomics or proteomics. In these fields, large correlational studies powered by machine learning have helped reveal new relationships. Genome-wide association studies, for example, can pick out possible links between complex genetic conditions and multiple genes that may contribute to the conditions.
High-throughput chemistry has generated large datasets of reaction outcomes at different temperatures using different catalysts and reagents, but big-data-style approaches haven’t been used much in this area so far, Lee says. As the cost of high-throughput chemistry drops and the data becomes available, though, that is changing.
“You’re...
Would you like to access the complete CEP News Update?
No problem. You just have to complete the following steps.
You have completed 0 of 2 steps.
-
Log in
You must be logged in to view this content. Log in now.
-
AIChE Membership
You must be an AIChE member to view this article. Join now.
Copyright Permissions
Would you like to reuse content from CEP Magazine? It’s easy to request permission to reuse content. Simply click here to connect instantly to licensing services, where you can choose from a list of options regarding how you would like to reuse the desired content and complete the transaction.
