Detection and identification of bacteria in a juice matrix with Fourier transform-near infrared spectroscopy and multivariate analysis

The use of Fourier transform-near infrared (FT-NIR) spectroscopy combined with multivariate pattern recognition techniques was evaluated to address the need for a fast and sensitive method for the detection of bacterial contamination in liquids. The complex cellular composition of bacteria produces FT-NIR vibrational transitions (overtone and combination bands), forming the basis for identification and subtyping. A database including strains of Escherichia coli, Pseudomonas aeruginosa, Bacillus subtilis, Bacillus cereus, and Bacillus thuringiensis was built, with special-care taken to optimize sample preparation. The bacterial cells were treated with 70% (vol/vol) ethanol to enhance safe handling of pathogenic strains and then concentrated on an aluminum oxide membrane to obtain a thin bacterial film. This simple membrane filtration procedure generated reproducible FT-NIR spectra that allowed for the rapid discrimination among closely related strains. Principal component analysis and soft independent modeling of class analogy of transformed spectra in the region 5,100 to 4,400 cm(-1) were able to discriminate between bacterial species. Spectroscopic analysis of apple juices inoculated with different strains of E. coli at approximately 10(5) CFU/ml showed that FT-NIR spectral features are consistent with bacterial contamination and soft independent modeling of class analogy correctly predicted the identity of the contaminant as strains of E. coli. FT-NIR in conjunction with multivariate techniques can be used for the rapid and accurate evaluation of potential bacterial contamination in liquids with minimal sample manipulation, and hence limited exposure of the laboratory worker to the agents.