Two commercial kits for identifying clinical yeast isolates, Analytab API 20C and BioMerieux Vitek Yeast ID 32C strips, and a simplified identification method (SIM) consisting of selected conventional tests and dedicated to foodborne yeasts, were evaluated for their performance in the identification of 166 yeast isolates from concentrated fruit juices and beverages. Twenty-five species representing 15 genera were identified. The percentage of isolates correctly identified by SIM was 91%, whereas the API 20C and ID 32C systems correctly identified 86 and 76%, respectively. The number of species not included in databases was 4 in SIM, 10 in API 20C, and 7 in ID 32C. Discrepant test reactions and errors in databases were responsible for failed and false identification. Nevertheless, the accuracy of individual test reactions was high, giving only 1.6% (SIM), 2.5% (API 20C), and 1.7% (ID 32C) discrepant results. Both commercial kits, if supplemented with a few tests necessary for identification of yeasts, can be easily applied to the SIM database developed for foodborne yeasts, thus offering a convenient and accurate identification procedure.
1Permanent address: Department of Microbiology, University of Horticulture and Food Industry, Budapest, Hungary.