Old and new concepts of species differentiation in Aspergillus.

The classification of the genus Aspergillus has been studied by many taxonomists. The most important monograph on which most taxonomies are derived from is strictly based on phenotypical characters. Later revisions of certain Aspergillus sections have been predominantly nomenclatural changes and primarily used morphological criteria. Many new taxa were added particularly in the genera Emericella and Neosartorya. Identification of the most common and often important species remains problematic due to the variability in the phenotypic characters. This has caused errors in the literature, especially concerning the links to mycotoxin formation. The new taxonomies are based on a polyphasic approach using phenotypical characters together with multigene DNA sequences. In a polyphasic approach micro- and macromorphology, physiology, metabolites produced and molecular data are all important, and in principle no particular method should be overemphasized. In particular extrolite profiles have proven to be specific for the taxa and this has contributed to a stable species concept, but DNA sequence data have also been very valuable in critical revisions of species and their taxonomy and phylogeny. Examples of new classifications for species in section Circumdati, Flavi,Fumigati and Nigri are presented. Although the polyphasic approach might reveal clear cut species, problems may arise for some species if they are to be separated based only on their microscopic features and few physiological features. Suggestions for new methods in order to carry out more fast and precise identifications will be discussed. Full genome sequencing and DNA arrays offers exciting new bases for identifying the Aspergilli, but recent methods based on image analysis of accurately fingerprinted phenotypes are also very promising. However both methods require a stable and well resolved taxonomy and nomenclature. Validated careful phenotypic classification (taxonomy) together with phylogenetic treatment of DNA sequence data is a prerequisite for reliable rapid identification methods and database formation. Concerning identification, DNA bar coding will be possible in the future, either based on molecular methods or certain phenotypic features.