In the battle against the pathogens that lurk on the carcasses of the animals we eat, sometimes the most difficult task is identifying which tiny enemy you are fighting. Scientists at the US Department of Agriculture’s Agricultural Research Service have developed a technique that allows them to identify and isolate one major bacterial foe, Campylobacter, within 24 hours. The novel approach uses hyperspectral imaging to distinguish Campylobacter targets from similar but nonthreatening bacteria. Shaped like puffy spiral staircases, members of the Campylobacter family are a major cause of food-borne illness. They typically flourish in the guts and waste of chickens, sheep, cattle and other animals, but can find themselves in the bellies of unsuspecting people if the animals’ carcasses aren’t processed and cooked appropriately. With millions of animals processed each year for human consumption, the USDA and other regulatory bodies, as well as food producers themselves, need a fast and accurate method for detecting Campylobacter. Investigators at the USDA’s Agricultural Research Service are using hyperspectral imaging to identify and isolate Campylobacter pathogens more rapidly than standard plating techniques can. Courtesy of Seung-Chul Yoon. The standard screening method for food-borne pathogens is to swab the animal, add the collected residue to a petri dish prepared in a growth medium such as agar, and then incubate the sample for up to 48 hours. Once bacterial colonies have been established, someone must visually ascertain whether the result is Campylobacter or something that looks like it. This adds time to the screening process. At the Agricultural Research Service, Seung-Chul Yoon and his colleagues have been working on hyperspectral imaging methods since the 1990s. Their latest technique captures spectral information across a wide swath of wavelengths for every pixel on an imaging array aimed at a target. In Yoon’s lab, the target was petri dishes loaded with pure samples of Campylobacter along with other microorganisms typically found on poultry carcasses. The group’s hypothesis was that Campylobacter and various other bacteria have just enough chemical differences that could be picked out by reading their spectral signatures. The researchers used a hyperspectral imaging system made by the Institute for Technology Development at the John C. Stennis Space Center in Mississippi. The CCD-based system features a diffraction grating spectrograph manufactured by Specim Spectral Imaging Ltd. of Oulu, Finland. The group used a pair of 50-W tungsten halogen lamps to light the samples, acquiring line scans of each sample; the diffraction grating enabled them to acquire 256 spectra in the range of 400 to 900 nm – at each of 640 pixels. They found that the optimum wavelength to differentiate Campylobacter from unrelated species was 503 nm when the bacteria were grown in blood agar and 501 nm when they were grown in Campy-Cefex agar. The technique worked with up to 99 percent accuracy, even when cultures were incubated for only 24 hours. They reported their efforts in the journal Sensing and Instrumentation for Food Quality and Safety in March 2010. According to Yoon, whereas scanning in the 400- to 900-nm range was enough to identify Campylobacter from among other bacteria, it could not differentiate one subspecies of Campylobacter from another. The next steps the group will take are designed to increase the level of difficulty to approach real-world conditions. “Our main efforts will be to establish a hyperspectral imaging technique [that is] reliably sensitive and specific for mixed cultures and carcass rinses with the 24-hour time window,” Yoon said. He added that the technique also might be used to screen such pathogens as salmonella and Escherichia coli.