Over the last ten years an explosive growth in the development of computer hardware has been matched by the appearance of many commercial and academic molecular modeling and structure-activity relationship (SAR and QSAR) packages. (The “Q" is used in QSAR when describing the structure of the compound in a quantitative way, the simplest examples of quantitative descriptors being the mass of the compound or the number of atoms present.) The molecular modeling and SAR packages have been primarily developed within the arena of the drug and agrochemical industries in an attempt to ease the search for relationships between structure and activity. The fragrance chemist has followed in the footsteps of the drug designer and used similar techniques in his own SAR studies. However, there have been very few studies comparing the usefulness of different SAR approaches to the field of olfaction. In addition, the rapid developments currently being made within the SAR field mean there is a continuing need to evaluate the use of new molecular modeling and SAR techniques in the study of structure-odor relationships.
This article describes the evaluation of three QSAR approacbes (CoMFA, Hansch and Principal Component Analysis) which were used to investigate the correlation between chemical structure and the fruitiness of esters. It includes the first published study of the use of comparative molecular field analysis (CoMFA) in the formulation of a structure-odor relationship.
Purpose of the Study
The objective of this work was to evaluate the relative usefulness of various QSAR approaches in understanding and predicting the odor properties of chemicals. The odor property chosen for this initial assessment was the perceived intensity of the fruity character of 27 aliphatic esters. This data set was considered to be ideal for the following reasons: