100A-27 |
Application of aluminum oxide membranes for the rapid classification of bacterial strains using Fourier Transform Near-Infrared spectroscopy and multivariate analysis |
L. E. RODRIGUEZ-SAONA1, F. S. Fry, Jr.2, F. M. Khambaty2, and E. M. Calvey2. (1) Joint Institute for Food Safety and Applied Nutrition (JIFSAN), University of Maryland and Food and Drug Administration, 5100 Paint Branch Parkway, College Park, MD 20740, (2) Center for Food Safety and Applied Nutrition, Food and Drug Administration, 5100 Paint Branch Parkway, College Park, MD 20740 Recent advances in Fourier transform near-infrared (FT-NIR) spectroscopic instrumentation and pattern recognition techniques have indicated a significant potential for monitoring the presence of and identifying microbial pathogens. The complex cellular composition of bacteria yields FT-NIR vibrational transitions that might be useful for identification and sub-typing. NIR absorption spectroscopy allows fast, accurate and non-destructive measurements of chemical components and can provide information about structural and physical properties of materials. The objective was to develop sample protocols for the reliable and reproducible identification and classification of bacterial strains by combining FT-NIR spectroscopy with multivariate methods. Bacteria, including multiple strains of Escherichia coli, Pseudomonas aeruginosa, Bacillus spp. and Listeria innocua were evaluated. The bacterial cells were treated with ethanol (70% v/v) to address safety concerns, and concentrated on an aluminum oxide membrane to obtain a thin bacterial film. FT-NIR measurements were made by using diffuse reflectance and the spectra were analyzed by Principal Component Analysis (PCA) and Soft Independent Modeling Class Analogy (SIMCA). A simple membrane filtration procedure yields a thin bacterial film resulting in increased sensitivity and allowing for the rapid discrimination among closely related bacterial strains. PCA and SIMCA of second derivative spectra in the 5100-4200 cm-1 region exhibited clusters that discriminated between bacteria species at levels ~0.5 mg wet cells weight (~ 106 CFU/mg). Factors such as film thickness and stage of growth substantially affected the FT-NIR spectra and diminished the ability of PCA to differentiate among strains; this underscores the importance of developing robust sampling protocols. 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. By generating a library of major foodborne pathogens and the refinement of the models, this approach could become a powerful tool for monitoring the safety of our food supply.
Session 100A, Food Microbiology: General II
|