Asymmetric catalysis, where one enantiomer is made preferentially over the other, is important in synthetic chemistry and in the creation of drug candidates. The development of a new reaction is difficult enough but predicting whether a catalyst will work on an untested substrate is even more so. Matthew Sigman and colleagues report how to train statistical models on existing reaction data from the chemical literature in order to understand the origin of asymmetric catalysis. As they demonstrate, using an imine addition reaction that makes chiral amines, the various parameters of each component can be understood and the mechanistic information can be used to design reactions. In this way, the enantioselectivity of a reaction can be predicted by their model with high precision.
Recent Hot Topics
Sign up for Nature Research e-alerts to get the lastest research in your inbox every week.