CLASSIFICATION OF SPECIES IN THE GENUS PENICILLIUM BY CURIE-POINT PYROLYSIS MASS-SPECTROMETRY FOLLOWED BY MULTIVARIATE-ANALYSIS AND ARTIFICIAL NEURAL NETWORKS

Curie point pyrolysis/mass spectrometry of Penicillium species was per formed with 530 degrees C Curie point foils. The mass spectra were sub mitted to principal component analysis, canonical variates analysis an d hierarchical cluster analysis, producing a final dendrogram by the u se of average linkage clustering. By this approach a successful classi fication of the species Penicillium italicum, P. expansum and P. digit atum originating from fruits was obtained. Isolates of the same specie s grouped together in the dendrogram, while the different species were distinguished. Also when grown on two different agar media, replicate s of the same species grouped together. Likewise, a satisfactory class ification was achieved by multivariate analysis of the data for variou s isolates of the cheese-associated fungi Aspergillus versicolor, P. d iscolor, P. roqueforti, P. solitum, P. verrucosum, P. commune and P. p alitans. However, some difficulties appeared in distinguishing the clo sely related species P. commune and P. palitans. Such difficulties bec ame greater on including more isolates and limiting the analysis to fi ve of the species. The use of back-propagation artificial neural netwo rks, in contrast, resulted in a correct classification in all cases. T hus, it is concluded that Curie point pyrolysis/mass spectrometry is u seful in chemotaxonomic studies of the closely related species in the genus Penicillium.