Accurate prediction of chemical reactions in solution is challenging for current state-of-the-art approaches based on transition state modelling with density functional theory. Models based on machine learning have emerged as a promising alternative to address these problems, but these models currently lack the precision to give crucial information on the magnitude of barrier heights, influence of solvents and catalysts and extent of regio- and chemoselectivity. Here, we construct hybrid models which combine the traditional transition state modelling and machine learning to accurately predict reaction barriers. We train a Gaussian Process Regression model to reproduce high-quality experimental kinetic data for the nucleophilic aromatic substitution reaction and use it to predict barriers with a mean absolute error of 0.77 kcal mol−1 for an external test set. The model was further validated on regio- and chemoselectivity prediction on patent reaction data and achieved a competitive top-1 accuracy of 86%, despite not being trained explicitly for this task. Importantly, the model gives error bars for its predictions that can be used for risk assessment by the end user. Hybrid models emerge as the preferred alternative for accurate reaction prediction in the very common low-data situation where only 100–150 rate constants are available for a reaction class. With recent advances in deep learning for quickly predicting barriers and transition state geometries from density functional theory, we envision that hybrid models will soon become a standard alternative to complement current machine learning approaches based on ground-state physical organic descriptors or structural information such as molecular graphs or fingerprints.
1 July 1998 in Journal of Computer-Aided Molecular Design
The enantioselective Pd-catalyzed redox-relay Heck arylation of acyclic alkenyl alcohols allows access to various useful chiral building blocks from simple olefinic substrates. Mechanistically, after the initial migratory insertion, a succession of β-hydride elimination and migratory insertion steps yields a saturated carbonyl product instead of the more general Heck product, an unsaturated alcohol. Here, we investigate the reaction mechanism, including the relay function, yielding the final carbonyl group transformation. M06 calculations predict a ΔΔG⧧ of 1 kcal/mol for the site selectivity and 2.5 kcal/mol for the enantioselectivity, in quantitative agreement with experimental results. The site selectivity is controlled by a remote electronic effect, where the developing polarization of the alkene in the migratory insertion transition state is stabilized by the C–O dipole of the alcohol moiety. The enantioselectivity is controlled by steric repulsion between the oxazoline substituent and the alcohol-bearing alkene substituent. The relay efficiency is due to an unusually smooth potential energy surface without high barriers, where the hydroxyalkyl-palladium species acts as a thermodynamic sink, driving the reaction toward the carbonyl product. Computational predictions of the relative reactivity and selectivity of the double bond isomers are validated experimentally.
5 May 2006 in Angewandte Chemie International Edition