Identification of Bacteria by Computer: Identification of Reference Strains

Abstract

SUMMARY: The results of the identification of 1079 reference strains of Gram-negative, aerobic, rod-shaped bacteria by a probabilistic method in a computer are given. Comparison of identification by conventional methods and by computer showed that 90.8% of fermentative and 82.1% of non-fermentative strains could be identified on the best available probability matrix. Many of these strains were atypical and had caused difficulty in identification in the medical diagnostic laboratory. The implications of various factors in successful computer identification are discussed and the results are given for each taxon by genus and species.