QSAR modeling of alpha-campholenic derivatives with sandalwood odor

Three-dimensional quantitative structure-activity relationship (3D-QSAR) models were developed for a series of 44 synthetic alpha-campholenic derivatives with sandalwood odor. These compounds have complex stereochemistry as they contain up to five chiral atoms. To address stereospecificity of odor intensity, a 3D-QSAR method was developed, which does not require spatial alignment of molecules. In this method, compounds are represented as derivatives of several common structural templates with several substituents, which are numbered according to their relative spatial positions in the molecule. Both wholistic and substituent descriptors calculated with the TSAR software were used as independent variables. Based on published experimental data of sandalwood odor intensities, two discrete scales of the odor intensity with equal or unequal intervals between the threshold values were developed. The data set was divided into a training set of 38 compounds and a test set of six compounds. To build QSAR models, a stepwise multiple linear regression method was used. The best model was obtained using the unequal scale of odor intensity: for the training set, the leave one out cross-validated R(2) (q(2)) was 0.80, the correlation coefficient R between actual and predicted odor intensities was 0.93, and the correlation coefficient for the test set was 0.95. The QSAR models developed in this study contribute to the better understanding of structural, electronic, and lipophilic properties responsible for sandalwood odor. Furthermore, the QSAR approach reported herein can be applied to other data sets that include compounds with complex stereochemistry.