The measurement of the responses to different odour intensities of 'boar taint' using a sensory panel and an electronic nose

This study explored the possibility of using an electronic nose (e-nose) with a 12-conducting-polymer sensor array combined with pattern recognition routines to discriminate between varying intensities of boar taint. A set of samples in a model system comprising a neutral lipid base containing varying combinations of androstenone and skatole were tested, as were pork fat samples. The e-nose responses for pork fat were calibrated against those given by a trained 10-member sensory panel for abnormal odour of the same samples from a total of 60 Large White cross-bred pigs. The e-nose responses related strongly to those of the sensory panel with a significant (p < 0.01) canonical correlation of 0.78. The data set was used to develop a discriminant function for grouping pork samples into three 'response classes': normal, doubtful and abnormal. Based on this, the e-nose identified all the abnormal samples correctly. However, 16% of the normal samples were also classified as abnormal. It was concluded that, in general, the electronic nose can discriminate between different levels of boar taint and that although a high specificity of sensors to androstenone and skatole may be desirable it may not be entirely important to the development and configuration of a boar taint sensor array. (C) 1998 Elsevier Science Ltd. All rights reserved