Pyrolysis-mass spectrometry characterisation of ovine fat tissues according to diet

The lipids of subcutaneous fat tissues from 120 lambs were analysed by Curi e point Pyrolysis - Mass Spectrometry (Py-MS) after extraction of the lipid s with hexane. The animals came from six European countries with different production systems. The classification of the fat tissues according to the predominant type of feed the lambs consumed (milk, concentrate or pasture) was obtained by Discriminant Analysis (DAs) and Artificial Neural Networks (ANNs). The classification models were validated by "leave-one-out"-type cr oss-validations. The results showed that the Py-MS spectra of the fat tissu e lipids permitted recognition of feed type. The percentages of correct cla ssifications determined by cross-validation were 78 % by DA and 92 % by ANN s.