<OL> <LI> To develop a rapid, automated FTIR-based system for identification of bacteria isolated from food.
<LI> To establish a comprehensive database comprising the spectra of microorganisms of food safety concern.
<LI> To develop an expert system encompassing appropriate data analysis algorithms for automated identification of unknowns.
<LI> To develop a sampling methodology for high-throughput screening, entailing the imprinting of bacteria grown on a hydrophobic grid membrane filter (HGMF) onto an IR-transparent optical window.
<LI> To evaluate and refinement of the FTIR method by testing in multiple laboratories.
<LI> Extensive validation to provide data required in order for the FTIR methodology to gain acceptance as a reliable alternative to well-established techniques of food microbiological analysis.
The benefits of this FTIR methodology would include more rapid analyses, the elimination of the need for reagents, a decrease in the cost of analysis per sample by an order of magnitude, and amenability to automation. <P>
Summary of Research Results:
The research in this project was undertaken in collaboration with researchers at the Health Canada Food Directorate to address the current requirements for new rapid, cost-effective methods for the identification of bacteria isolated from food and water. The project concerned bacteria identification based on whole-organism fingerprinting by Fourier transform infrared (FTIR) spectroscopy. The FTIR spectra of ~150 strains of foodborne bacteria, belonging to a wide variety of species and fully characterized by standard microbiological methods, were recorded following growth of the bacteria on an \"IR-grade\" universal growth medium. This growth medium was found to support the growth of all the bacterial strains tested throughout the course of the project, allowing a standard sample preparation protocol to be established. An automated x-y stage was constructed to allow for sequential acquisition of the spectra of ~200 samples deposited on an infrared-transparent slide. The use of FTIR imaging instrumentation provided an effective means of compensating for the effects of sample heterogeneity. A hierarchical classification strategy and a region selection algorithm that was developed for this project resulted in differentiation between all the strains in the database from one another based on differences in their FTIR spectra. Validation studies entailed re-growing all the strains that had been included in the database, recording their spectra, and identifying them as \"unknowns\". All strains were correctly identified at the genus and species levels. Based on analysis of the data for a subset comprising 44 strains of Clostridium botulinum, FTIR spectroscopy was shown to have a discriminatory power comparable to that of pulsed-field gel electrophoresis, a genotypic method that is currently the \"gold standard\" for typing of foodborne bacterial pathogens. FTIR spectrometers were installed in two Health Canada Food Directorate laboratories for validation and evaluation of the FTIR bacteria identification methodology and the spectral database of foodborne bacteria. The technology developed in this project was protected in a patent application filed by the Office of Technology Transfer of McGill University. The University has also entered into negotiations with a major instrument manufacturer to license the software written for FTIR bacteria identification. This work was made possible by the exceptional dedication of Mr. Jonah Prevost Kirkwood and Mr. Andrew Ghetler, graduate students in the Department of Food Science and Agricultural Chemistry of McGill University., and the expertise provided by Dr. John Austin, Mme. Lorraine Gour and Mme. Irene Iugovazof the Food Directorate of Health Canada. <P> For more information, please visit the <a href="http://www.omafra.gov.on.ca/english/research/foodsafety/index.html" target="_blank">Ontario Ministry of Agriculture, Food & Rural Affairs (OMAFRA) Food Safety Research Program</a>.