Guiding Chemical Synthesis: Computational Prediction of the Regioselectivity of CH Functionalization

Abstract

We will develop a computational method (RegioSQM) for predicting the regioselectivity of CH functionalization reactions that can be used by synthetic chemists who are not experts in computational chemistry through a simple web interface (regiosqm.org). CH functionalization, i.e. replacing the hydrogen atom in a CH bond with another atom or molecule, is arguably single most promising technique for increasing the efficiency of chemical synthesis, but there are no generally applicable predictive tools that predict which CH bond is most reactive. RegioSQM uses semiempirical quantum chemistry methods to predict relative stabilities of reaction intermediates which correlates with reaction rate and our goal is to determine which quantum method and intermediate give the best result for each reaction type. Finally, we will experimentally demonstrate how RegioSQM can be used to make the chemical synthesis part of drug discovery more efficient.