Hybrid Machine Learning Approach to Predict the Site Selectivity of Iridium-Catalyzed Arene Borylation

The borylation of aryl and heteroaryl C-H bondsis valuablefor the site-selective functionalization of C-H bonds in complexmolecules. Iridium catalysts ligated by bipyridine ligands catalyzethe borylation of the C-H bond that is most acidic and leaststerically hindered in an arene, but predicting the site of borylationin molecules containing multiple arenes is difficult. To address thischallenge, we report a hybrid computational model that predicts theSite of Borylation (SoBo) in complex molecules. The SoBo model combinesdensity functional theory, semiempirical quantum mechanics, cheminformatics,linear regression, and machine learning to predict site selectivityand to extrapolate these predictions to new chemical space. Experimentalvalidation of SoBo showed that the model predicts the major site ofborylation of pharmaceutical intermediates with higher accuracy thanprior machine-learning models or human experts, demonstrating thatSoBo will be useful to guide experiments for the borylation of specificC(sp(2))-H bonds during pharmaceutical development.