Discovery of Novel Glycosylation Methods using Bayesian Optimization: Lithium Salt Directed Stereoselective Glycosylations

Abstract

In recent years, Bayesian optimization hasgained increasing interest as a tool for reaction optimization.Here we use Bayesian optimization in a reaction discoveryfashion by treating the glycosylation reaction classas a black box function. This provides access to new areasof the glycosylation reaction space and leads to the discoveryof novel stereoselective glycosylation methodologies,where stereoselectivity can be directed by the addition oflithium salts in interplay with other reaction conditions.Black box functions are inherently difficult to interpret,but we show how partial dependence plots can be used toinfer trends from the obtained data in a similar fashion tothe commonly used one-variable-at-time approach.