In Silico Prediction of Ionization

Authors: Fraczkiewicz R
Publication: Comprehensive Medicinal Chemistry II, Vol. 5
Software: ADMET Predictor®
Division: Simulations Plus

Abstract

The original version of this chapter was written for the edition of “Comprehensive Medicinal Chemistry II”1 published by Elsevier in 2006. This version has been revised, expanded, and updated to accommodate new methodological advancements in pKa prediction that have taken place between 2006 and 2012.The focus of this chapter is computational methods for the prediction of aqueous ionization of organic molecules. The thermodynamic background, as well as molecular factors influencing ionization, is discussed in detail. The microscopic view of ionization is explained. Methods of prediction are reviewed with particular attention to developments of the last 30 years. Purely theoretical and empirical approaches are identified. The latter are divided into two major types: 1) classic database lookup+perturbational LFER (linear free energy relationships) and 2) QSPR (quantitative structure-property relationships). QSPR methods are considered from purely empirical and quantum mechanical points of view. In addition, ab initio quantum mechanical approaches to ionization of small molecules and statistical thermodynamic treatment of the ionization of proteins and nucleic acids are described. The newly added following section comments on experimental pKa data and specifies rules for objective evaluation of predictive models of pKa. Available software packages and sources of ionization data are reported next. An outlook on the future of ionization prediction closes this chapter.
By Robert Fraczkiewicz