As part of the development of our in-house molecular modelling package, COSMIC, the structure-property relationships of a series of 13 antimalarial chloroquine derivatives have been investigated using physicochemical properties calculated from molecular models. This has involved the use of various statistical techniques. The compounds were analysed using three different modelling approximations concerning assumptions about the invariant amino alkyl side chain. The first two sets of data were obtained from models with and without the side shain using a formal charge of 0.0. Analysis of these data using principal components resulted in plots of principal component scores in which activity categories were separated. In the third data set, the side chain was modelled with a charge of +1 and an examination of the molecular properties showed poorer clustering of activity categories. We have also investigated relationships between biological activity and physicochemical properties using multiple linear regression. Although significant equations were developed, the correlation coefficients were low and it was not felt that they would be useful for quantitative prediction. The identification of important variables, however, may give some clues to the mechanism of action of these compounds.
- Antimalarial drugs
- Multiple Linear Regression
- Multivariate Analysis
- Principal Components Analysis
- Quantitative Structure-Activity Relationships