|
UK scientists have developed a technology that can identify contaminated meat in seconds using infrared spectroscopy and genetic algorithms. The researchers reported their latest results on 14th September at the Society for General Microbiology meeting at Keele University.
Meat spoils because of the growth and enzymatic activity of microorganisms leading to decomposition and the formation of unpleasant and odorous metabolites. There are already more than forty different tests for meat spoilage but most of these involve taking samples and separating bacteria from the meat are not as quick as the modern food industry requires.
David Ellis and his colleagues David Broadhurst, Douglas Kell, Jem Rowland, and Roy Goodacre of the University of Manchester have for several years focused on the use of Fourier transform infrared spectroscopy as a rapid, non-invasive, and reagent-free method of spotting spoilage from the tell-tale IR fingerprints of food spoilage bacteria. Ellis suggests that spectroscopy coupled with a computer-based statistical interpretation method meets the requirements for a rapid, non-invasive, reagent free and relatively inexpensive approach to testing meat.
He and his colleagues have now demonstrated that FTIR spectroscopy can be used directly on the surface of food to produce biochemically interpretable "fingerprints". These fingerprints are interpreted using partial least-squares regression and allowed the researchers to estimate accurately the number of bacteria in a sample.
The researchers exploited genetic programming, a programming technique based on the principles of "survival of the fittest" that leads to the best program for performing a particular task by excluding the failures and building the next generation on the basis of the successes. Genetic programming helped them derive a set of rules that would reveal the onset of spoilage on the basis of the input FTIR spectra. "Using FTIR and machine learning we were able to acquire a metabolic snapshot and quantify, non-invasively, the microbial loads of food samples accurately and rapidly in 60 s, directly from the sample surface," Ellis said.
The researchers have now demonstrated efficacy on beef, as well as poultry. Beef muscle is a more complex substrate, so their success suggests that less complex foodstuffs, such as dairy produce, should also be amenable to the test. "We can detect levels of bacteria well below the level when they are detectable by the human nose," Ellis told SN. Indeed, he adds, the food industry will recognize the benefits as the technique can detect sensory spoilage in chicken, for example, two log values below the olfactory limit and even lower for beef. The main advantage, however, is that their method is extremely rapid whereas current tests take hours to carry out.
"The process could easily be miniaturized and incorporated either on-line or just off-line of the production process," Ellis told us, "and indeed such technology already exists in the form of hand-held and portable IR spectrometers."
Related Links: |
Pureed chicken - ready for analysis.
FTIR analyser
|
