Automated, rapid identification of bacteria by pattern analysis of growth inhibition profiles obtained with Autobac 1

Abstract

A scheme for identifying bacteria has been devised which utilizes the inhibition patterns obtained by Autobac 1 with routine and unusual antimicrobial agents and with other differentially inhibitory chemical compounds. Over 600 compounds were initially identified from the literature, and over 125 of these were selected for further testing on the basis of antibacterial activity most conducive to the instrument-generated differential scheme. Numerical growth index information derived by light scatter comparisons from the instrument were analyzed by computer, utilizing the quadratic discriminant function statistical technique. In comparison with conventional methods, accuracy for the 10 bacterial genera studied was 95% or greater. Results indicate a potential for both bacterial identification and antimicrobial agent susceptibility testing in the clinical laboratory within 3 to 5 h when using this automated approach.