2012 Annual Report
Once reasonable results are obtained for pure cultures, important food matrices will be spiked with pathogens and bio-threat agents at varying concentration levels and hyperspectral images will be collected on both the total contaminated food matrices and on a rinsate from the infected food matrices. Multivariate calibration models, based on the spectral libraries of the bio-threat agents and food matrices, will then be developed and the lower limit of detection for each pathogen and bio-threat agent will be determined.
Previous work described the creation of spectral libraries of pure Shiga-Toxin producing Escherichia Coli (STEC) serogroups other than E. coli O157:H7 (non-O157) which are better known as non-O157 STEC. These spectral libraries were then used for classification model development with hyperspectral imaging. In this study, we compared the spectra of pure non-O157:H7 STEC serogroups to the spectra of non-O157:H7 STEC serogroups contaminated (i.e. spiked) in ground beef. The mean spectra of pure non-O157:H7 STEC serogroups presented earlier and the same serogroups spiked into the broth of ground beef had similar spectral responses. Serogroup O111 also showed a distinctive spectral signature linearly increasing from 500 to 650 nm and was less absorptive in the blue and green portions of the spectrum (esp. 475-575 nm) than all the other serogroups. Serogroups O26, O103, and O145 showed similar spectral responses with purple and/or gray-blue centers surrounded by more reddish perimeters. Serogroup O45 was highly absorptive across the visible spectrum, with the 500 to 700 nm regions demonstrating the greatest absorption, while the 400 to 500 nm region (blue) was somewhat less absorptive. This resulted in a dark blue appearance to the human eye and rendered it spectrally unique relative to the other serogroups.
The pair-wise comparisons of each pure serogroup versus the spiked serogroups showed difference in the spectra of pure O26 versus spiked O26. The average pure spectra spiked O103 and O111 were similar but the pure spectra were slightly more absorptive (less reflectivity) than the spiked spectra. However, the greatest difference between pure and spiked spectra was for serogroup O45. In general, the pure O45 had a higher absorption than the spiked O45 across the entire spectral range, with 500 to 700 nm region demonstrating the greatest difference. This results in the pure O45 having a darker blue appearance to the human eye than the spiked O45 colonies. Nevertheless, it is still spectrally unique relative to the other serogroups. Futhermore, spectral libraries and classification models developed from the pure serogroups are applicable to detect and classify the serogroups in ground beef and also in the presence of background flora.
Earlier research reported the development of classification models for pure Bacillus anthracis Stern from spectral libraries of hyperspectral imaging data. In this research, the spectra of pure B. anthracis Stern was compared to the spectra of B. anthracis Stern contaminated (i.e. spiked) in hot dog. The hotdog samples spiked with B. anthracis Stern were plated on sheeps blood agar (SBA) and incubated for 15 to 24 hours without any additional background microflora. There was a replicate affect do to different types of hotdogs tested but no difference in the B. anthracis Stern spectrum.
Salmonella Typhimurium (ST) was spiked in poultry carcass rinses and plated on brilliant green sulfa (BGS) agar with and without Nalidixic acid. In general, ST colonies grew with reddish color and circular shape in BGS agar. The spectral signatures of the hyperspectral data from pure ST were similar to each other between samples from the plates with and without Nalidixic acid. However, some other non-ST colonies had similar color and shape to ST. These colonies had spectral signatures that increased at lower wavelengths but then decrease between 400 nm and 560 nm (blue and green color bands). Confirmation of each colony type is currently being conducted with a real-time polymerase chain reaction (PCR). Similar studies were conducted with ST growing on Xylose Lysine Terigtol 4 (XLT4) agar. In this research, it was difficult to discriminate the ST from the background microflora. In XLT4, colonies that changed to a black color are supposed to be ST. However, some black-colored colonies were identified as different pathogens than ST. Further experiments are being conducted to validate the results. In general, the green and red color bands from the colonies growing in XLT4 agar plates were more important in identifying Salmonella in the spectral domain. In addition to spectral information, different morphological characteristics (between circular and irregular shape) of the colonies growing on the agar plate need to be incorporated into future classification models.