Bacterial identification by near-infrared chemical imaging of food-specific cards

Near-infrared chemical imaging (NIR-CI) is investigated as a tool for the high-throughput analysis of self-contained microbial identification test cards for micro-organisms of concern in food. In this initial work, a NIR-CI system operating in the spectral range 1000-2350 nm was used to acquire NIR chemical images of bacterial cells deposited on a 'card', containing both the calibration and test samples. Results show that some bacteria can be identified from differences observed at unique wavelengths, and that a standard operating procedure can be developed for a particular 'card' to differentiate and hence identify the various organisms it contains using discrete wavelengths. For situations where a particular organism of concern is sought, a PLS chemometric model may offer better performance by accounting for variables that can be incorporated in the calibration without the need to know the taxonomic identity of the complete complement of bacteria present on the 'card'. Overall, the NIR-CI results obtained in this investigation show that this high throughput technique possesses the specificity required to differentiate bacteria on the basis of their NIR spectra.