Concept for Development of Expert Computer Program for Identification of Filamentous Fungi

Abstract

An expert program has been developed for users working in industrial laboratories who are not experts in the identification of filamentous fungi. The database of morphological growth features currently contains 12 species from the genera Aspergillus and Penicillium grown under standard conditions. The identification algorithm implemented in the database takes into account the reliability of users, which can vary over a wide range depending on the identification feature. The reliability of users was estimated on the basis of a questionnaire survey conducted among 27 non-experts, as the likelihood of a response consistent with the assessment of experts. The program works through comparative analysis of features of the fungus being identified with the expert-developed database and selection of the most likely species among the species represented by reference strains. The expert program reduces subjective mistakes and may be extended to include further fungal species and genera; it can also be supplemented with chemotaxonomic, genetic and other data.