Physicochemical computation

Karpov, I.K. 1981. The physicochemical computer modeling in geochemistry. Novosibirsk Nauka. 

The method that was developed builds on computed values of physicochemical effects and uses neural networks for classification. Therefore, for a deeper understanding of this form of reaction classification, later chapters should be consulted on topics such as methods for the calculation of physicochemical effects (Section 7.1) and artificial neural networks (Section 9.4). 

An extensive series of studies for the prediction of aqueous solubility has been reported in the literature, as summarized by Lipinski et al. [15] and jorgensen and Duffy [16]. These methods can be categorized into three types 1 correlation of solubility with experimentally determined physicochemical properties such as melting point and molecular volume 2) estimation of solubility by group contribution methods and 3) correlation of solubility with descriptors derived from the molecular structure by computational methods. The third approach has been proven to be particularly successful for the prediction of solubility because it does not need experimental descriptors and can therefore be applied to collections of virtual compounds also. 

K and G M Crippen 1986. Atomic Physicochemical Parameters for Three-dimensional Struc-directed Quantitative Structure-Activity Relationships. I. Partition Coefficients as a Measure ydrophobicity. Journal of Computational Chemistry 7 565-577. 

The holistic thermodynamic approach based on material (charge, concentration and electron) balances is a firm and valuable tool for a choice of the best a priori conditions of chemical analyses performed in electrolytic systems. Such an approach has been already presented in a series of papers issued in recent years, see [1-4] and references cited therein. In this communication, the approach will be exemplified with electrolytic systems, with special emphasis put on the complex systems where all particular types (acid-base, redox, complexation and precipitation) of chemical equilibria occur in parallel and/or sequentially. All attainable physicochemical knowledge can be involved in calculations and none simplifying assumptions are needed. All analytical prescriptions can be followed. The approach enables all possible (from thermodynamic viewpoint) reactions to be included and all effects resulting from activation barrier(s) and incomplete set of equilibrium data presumed can be tested. The problems involved are presented on some examples of analytical systems considered lately, concerning potentiometric titrations in complex titrand + titrant systems. All calculations were done with use of iterative computer programs MATLAB and DELPHI. 

To find the best a priori conditions of analysis, the equilibrium analysis, based on material balances and all physicochemical knowledge involved with an electrolytic system, has been done with use of iterative computer programs. The effects resulting from (a) a buffer chosen, (b) its concentration and (c) complexing properties, (d) pH value established were considered in simulated and experimental titrations. Further effects tested were tolerances in (e) volumes of titrants added in aliquots, (f) pre-assumed pH values on precision and accuracy of concentration measured from intersection of two segments obtained in such titrations. 

The partial differential equations used to model the dynamic behavior of physicochemical processes often exhibit complicated, non-recurrent dynamic behavior. Simple simulation is often not capable of correlating and interpreting such results. We present two illustrative cases in which the computation of unstable, saddle-type solutions and their stable and unstable manifolds is critical to the understanding of the system dynamics. Implementation characteristics of algorithms that perform such computations are also discussed. 

Purdy [91] used the technique to predict the carcinogenicity of organic chemicals in rodents, although his model was based on physicochemical and molecular orbital-based descriptors as well as on substructural features and it used only a relatively small number of compounds. His decision tree, which was manual rather than computer based, was trained on 306 compounds and tested on 301 different compounds it achieved 96% correct classification for the training set and 90% correct classification for the test set. 

This book focuses on molecular features and properhes, their meaning, measurement, computation, and encoding into parameters and descriptors. The present chapter serves as a general opening, and invites readers to stand back and reflect on the information contained in chemical compounds and on our descrip-hon of it. We base our approach on a discriminahon between the core features of a molecule/compound and the physicochemical properhes of a compound. 

The growing computahonal power available to researchers proves an invaluable tool to investigate the dynamic profile of molecules. Molecular dynamics (MD) and Monte Carlo (MC) simulahons have thus become pivotal techniques to explore the dynamic dimension of physicochemical properhes [1]. Furthermore, the powerful computational methods based in parhcular on MIFs [7-10] allow some physicochemical properhes to be computed for each conformer (e.g. virtual log P), suggesting that to the conformahonal space there must correspond a property space covering the ensemble of all possible conformer-dependent property values. 

Testa, B., Van de Waterbeemd, H., Folkers, G., Guy, R. (eds.). Pharmacokinetic Optimization in Drug Research Biological, Physicochemical and Computational Strategies, VHGA, Zurich and Wiley-VGH, Weinheim, 2001. 

---