RAPID IDENTIFICATION AND CLASSIFICATION OF STAPHYLOCOCCUS AUREUS BY ATTENUATED TOTAL REFLECTANCE FOURIER TRANSFORM INFRARED SPECTROSCOPY

摘要

Staphylococcus aureus is an important bacterium that can cause serious infections in human such as pneumonia and bacteremia. Rapid detection of this pathogen is crucial in food industries and clinical laboratories to control S. aureus food poisoning and human infections. In this study, Fourier transform infrared spectroscopy equipped with a germanium attenuated total reflection accessory was used as a novel approach to identify S. aureus. A total of 17 reference strains belonging to 4 different species and 84 clinical isolates of Staphylococcus spp. were analyzed. After the cultivation of the strains, spectral collection and data preprocessing, the S. aureus isolates were identified by a two-step discrimination procedure. An internal validation andthe related external validation were performed to demonstrate the discriminatory power and the quality of the discrimination models before the discrimination analysis. In the first step, 38 S. aureus isolates were correctly classified and the others weremisidentified as Staphylococcus haemolyticus by hierarchical clustering analysis model using the first derivatives from the spectral range between 1,800 and 1,050/cm. In the second step, several classification/discrimination algorithms of soft-independent modeling of class analogy, principal component regression and partial least squares regression (PLSR) were applied to build models for differentiating S. aureus and S. haemolyticus. The results showed that 57 (98.3%) strains and 4 (100%) strains of S. aureus and S. haemolyticus could be correctly identified by PLSR.