Quantitative structure activity relationships (QSAR) and comparative molecular field analysis (CoMFA) are applied in order to explain the aroma of 46 bell-pepper aroma compounds. Biological activities log(1/c) values are used, where c stands for the detection threshold value of the aroma compound in water. Results of conventional QSAR and CoMFA are both satisfactory in statistical significance and predictive ability. We construct a qualitative model using the graphic features of CoMFA together with the results of "classical" QSAR analysis, which is performed by multiple linear regression. Finally, the human olfactory detection threshold values of excluded pyrazines are successfully predicted. This makes CoMFA and QSAR two important tools for designing new aroma compounds and in elucidating the mechanism of odor-receptor interaction.

An artificial neural network is used to predict both the classification of aroma compounds and their flavor impression threshold values for a series of pyrazines. The classification set consists of 98 compounds (32 green, 43 bell-pepper, and 23 nutty smelling pyrazines), and the regression sets consist of 24 green and 37 bell-pepper odorous pyrazines. The best classification of the three aroma impressions (93.7%) is obtained by using a multilayer perceptron network architecture. To predict the threshold values of bell-pepper fragrance, a standard Pearson R correlation coefficient of 0.936 for the training set, 0.912 for the verification set, and 0.926 for the test set is received with two hidden layers consisting of two and one neurons. The network for the threshold prediction of the class of green-smelling pyrazines with one hidden layer containing three neurons turns out to be the best with a standard Pearson R correlation coefficient of 0.859 for the training, 0.918 for the verification, and 0.948 for the test set. These good correlations show that artificial neural networks are versatile tools for the classification of aroma compounds.

Furan derivatives are part of nearly all food aromas. They are mainly formed by thermal degradation of carbohydrates and ascorbic acid and from sugar-amino acid interactions during food processing. Caramel-like, sweet, fruity, nutty, meaty, and burnt odor impressions are associated with this class of compounds. In the presented work, structure-activity relationship (SAR) investigations are performed on a series of furan derivatives in order to find structural subunits, which are responsible for the particular characteristic flavors. Therefore, artificial neural networks are applied on a set of 35 furans with the aroma categories "meaty" or "fruity" to calculate a classification rule and class boundaries for these two aroma impressions. By training a multilayer perceptron network architecture with a backpropagation algorithm, a correct classification rate of 100% is obtained. The neural network is able to distinguish between the two studied groups by using the following significant descriptors as inputs: number of sulfur atoms, Looping Centric Information Index, Folding Degree Index and Petitjean Shape Indices. Finally, the results clearly demonstrate that artificial neural networks are successful tools to investigate non-linear qualitative structure-odor relationships of aroma compounds.

The encoding of various aroma impressions and the distinction between different aroma qualities are unsolved problems, as differences between aroma impressions can be described only in a qualitative but not in a quantitative manner. As a consequence, classifications of various aroma qualities cannot easily be performed by standard QSAR methods. To find a proper way to encode aroma impressions for SAR studies, a total of 50 pyrazine-based aroma compounds showing the aroma quality of earthy, green-earthy, or green are analyzed. Special attention is thereby turned on the mixed aroma impression green-earthy. Classifications on the whole data set as well as on smaller subsets are calculated using self-organizing molecular field analysis (SOMFA) and artificial neural networks (ANNs). SOMFA classifies between two or three aroma impressions, leading to models satisfying in predictive power. ANN analysis using multilayer perceptron network architecture with one hidden layer and nominal output as well as genetic regression neural network) with two hidden layers and numerical output both lead to a rather good performance rate of 94%.

No abstract