Toward a computational tool predicting the stereochemical outcome of asymmetric reactions: Development of the molecular mechanics‐based program ACE and application to asymmetric epoxidation reactions

Abstract

The development and application of ACE, a program that predicts the stereochemical outcome of asymmetric reactions is presented. As major implementations, ACE includes a genetic algorithm to carry out an efficient global conformational search combined with a conjugate gradient minimization routine for local optimization and a corner flap algorithm to search ring conformations. Further improvements have been made that enable ACE to generate Boltzmann populations of conformations, to investigate highly asynchronous reactions, to compute fluctuating partial atomic charges and solvation energy and to automatically construct reactants and products from libraries of catalysts and substrates. Validation on previously investigated reactions (asymmetric Diels Alder cycloadditions and organocatalyzed aldol reactions) followed by application to a number of alkene epoxidation reactions and a comparative study of DFT-derived and ACE-derived predictions demonstrate the accuracy and usefulness of ACE in the context of asymmetric catalyst design.

Regularized regression analysis of digitized molecular structures in organic reactions for quantification of steric effects

“…Recently, the regression analysis of catalytic asymmetric reactions has attracted attention, and several descriptors and protocols have been examined . CoMFA and related methods have also been used for the analysis of such transformations and have provided insights into reactions through the visualization of important molecular structural information .…”

Improving the Q2MM method for transition state force field modeling

“…Finally, we want to note that methods that allow force field determinations of TSs by mixing of reactants and products, like the empirical valence bond (EVB) model potential, SEAM, and ACE, could potentially determine TS structures without suffering the erroneous distortion shown in Figure . However, none of those methods have yet been tested with a parameterization that allows reproduction of the TS normal modes, and thus have not been able to achieve the accuracy that has been demonstrated with Q2MM.…”

Efficient Transition State Modeling Using Molecular Mechanics Force Fields for the Everyday Chemist

“…ACE 2.0 also approximates a Boltzmann population, designed to improve the prediction of temperature-dependent phenomena. 108 In a more recent version (ACE 2.2, unpublished), the focus was on automating the entire computation protocol. In ACE 2.0, each TS (two diastereomeric TSs in Figure 11b) was computed separately, and the predicted diastereomeric excesses (d.e.)…”

“…The generated TS structures were compared to those reported previously using DFT methods (Figure 12). 108 The predicted selectivities were compared to experiment and also to DFT-generated predictions. A good correlation was found with mean unsigned error in the range of 1 kcal/mol.…”