William J. Dunn

Chemicobiological interactions and the use of partition coefficients in their correlation

Abstract Using octanol/water partition coefficients as an operational definition of hydrophobicity, 70 examples are given in which the hydrophobic interactions of organic compounds with themselves (in micelles) with macromolecules or with biological systems can be quantitatively correlated by the expression: log RBR = a log P + b. In this expression RBR is a binding constant or a relative biological response, P is the octanol/water partition coefficient and a and b are constants obtained via the method of least squares. These results are strong support for the utility of log P in the correlation of hydrophobic interactions. They also illustrate the extremely wide range of processes in which hydrophobic bonding plays a critical role.

Relationships between chemical structure and biological activity modeled by SIMCA pattern recognition

Abstract For a class of chemically and pharmacologically similar compounds, one can formulate a quantitative relation between the variation in chemical structure and the variation in measured biological activity. This relation is based on (1) the translation of the compound structures into quantitative variables by means of either substituent parameters derived from the influence of the substituents on chemical model reactions or theoretical variables derived from quantum chemical or other theoretical calculations, (2) multivariate statistical models extracting the common “pattern” of the structural descriptor variables for the compounds in the class, and (3) relations between parameters emerging from the statistical models and the measured biological activities of the compounds. In cases when several classes of compounds are studied, the data analysis also involves a classification, and the total analysis becomes one of pattern recognition or discriminant analysis. The methodology is illustrated by means of four examples: the classification and activity prediction of some β-adrenergic compounds, the prediction of the carcinogenicity of some 4-nitroquinoline-1-oxides, the prediction of the carcinogenicity of some polycyclic aromatic hydrocarbons, and the prediction of the glycemic activity of some o -toluenesulfonyl (thio)ureas.

Clustering of aryl carbon-13 nuclear magnetic resonance substituent chemical shifts. A multivariate data analysis using principal components

A principal component analysis of the 13C substituent-induced chemical shifts of 82 monosubstituted benzenes shows that ca. 90% of the substituents belong to one of four clusters, acceptors, alkyls, donors, or halogens. This grouping is confirmed statistically. The extensions of the subclasses are not parallel. It is also shown that the predictive capability of the single-parameter models for each subclass is better than any multiparameter model applied on the whole data set. The observed grouping of substituents provides an explanation to the apparent correlation frequently found between 13C n.m.r. chemical shifts and dual substituent parameters. The ability of the statistical method to discover incorrect shift data is also illustrated.

An assessment of the carcinogenicity of N-nitroso compounds by the SIMCA method of pattern recognition

The ability to predict the toxic responses of potential environmental pollutants on the basis of their physiochemical properties has many advantages. Pattern recognition methods can be used to predict such pharmacological properties. In this report the SIMCA method of pattern recognition is used to predict the carcinogenicity of N-nitroso compounds, and the advantages of this method of pattern recognition in such applications are discussed.

The carcinogenicity of N-nitroso compounds: A SIMCA pattern recognition study

Abstract A number of methods are available for evaluating the potential of chemical agents to induce cancer in test animals. Recent use of SIMCA pattern recognition in structure-biological activity studies suggests that this method may be useful in making such evaluations. In this report this method is used to estimate the potential of a number of N-nitroso compounds to induce tumors in the rat. The data treatment and estimation results are consistent with the chemistry of these substances.