RESUMO
Identification of clinical yeast isolates causing candidiasis is routinely performed by commercial yeast identification systems based on biochemical, morphological and physiological tests. These systems require 3-5 days and the proportion of identifications that are incorrect is high. Our novel and rapid molecular identification system for clinical Candida species is based on the analysis of restriction patterns obtained from PCR-generated ribosomal DNA sequences using five restriction enzymes. A software package (CandID) was designed to include a database of restriction fragment length polymorphism (RFLP) patterns for 29 Candida species. For 'in-house' validation, 122 clinical isolates that had previously identified in clinical laboratories were typed by this system. These clinical isolates were also independently re-identified by the API 20C AUX system. The ribosomal DNA RFLP database in the context of supporting analytical software allowed simple and rapid (1 work day) identification.