Construction of a Statistical Evaluation Model Based on Molecular Centrality to Find Retrosynthetically Important Bonds in Organic Compounds

Abstract

For the purpose of finding retrosynthetically important bonds in a molecule, a new evaluation score has been defined through a logistic regression analysis of known reactions stored in reaction databases. We conceived that reaction center bonds in reaction databases describe one of the most retrosynthetically important bonds for each product structure. The derived statistical equation consists of bond centrality and bond dissociation energy terms. The equation shows that synthetically useful bonds tend to be more central in a molecule and to be weaker bonds. Coefficients in two statistical equations derived from two different reaction data sets are quite similar to each other. From a comparison of molecular complexities and validation with 35 complicated organic compounds, the evaluation equation was proved to be useful. (© Wiley‐VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2008)

Graph-Mining Algorithm for the Evaluation of Bond Formability

“…The latter can be perceived as an improvement of the former as it includes the molecular centrality and two additional parameters: the bond dissociation energy and the number of chiral centers. 36 From these three parameters and through a logistic regression analysis, authors introduce a new scoring function assessing the retrosynthetic importance of bonds by studying reaction centers from two available reaction databases. Though the results are rather satisfactory, only one bond is considered as the analysis output whereas synthetic chemists often consider several bonds as retrosynthetically interesting according to the molecule functionality.…”

“…The fact of being a strategic bond is rather a fuzzy notion that may lead to different formal definitions, and we focus on recent approaches dealing with retrosynthetically important bonds. Apart from Bertz’s and Hanessian’s studies cited in the introduction that deal with the topological and stereochemical strategic disconnections, two other formal approaches were recently developed to find retrosynthetically important bonds. , The former proposes a new index, called molecular centrality, based on the concept of convergent synthesis and computed from quadratic shortest path distances . This index is mainly topological and proves to be moderately useful in case of highly functionalized molecules.…”

“…This index is mainly topological and proves to be moderately useful in case of highly functionalized molecules. The latter can be perceived as an improvement of the former as it includes the molecular centrality and two additional parameters: the bond dissociation energy and the number of chiral centers . From these three parameters and through a logistic regression analysis, authors introduce a new scoring function assessing the retrosynthetic importance of bonds by studying reaction centers from two available reaction databases.…”