Skip to main content
ARS Home » Southeast Area » Athens, Georgia » U.S. National Poultry Research Center » Poultry Microbiological Safety and Processing Research Unit » Research » Publications at this Location » Publication #325423
Title: Detection and classification of salmonella serotypes using spectral signatures collected by fourier transform infrared (FT-IR) spectroscopy

Author
item SUNDARAM, JAYA - University Of Georgia
item Park, Bosoon
item Hinton Jr, Arthur

Submitted to: Proceedings of the American Society of Agricultural and Biological Engineers International (ASABE)
Publication Type: Abstract Only
Publication Acceptance Date: 2/8/2016
Publication Date: 7/18/2016
Citation: Sundaram, J., Park, B., Hinton Jr, A. 2016. Detection and classification of salmonella serotypes using spectral signatures collected by fourier transform infrared (FT-IR) spectroscopy. Proceedings of the American Society of Agricultural and Biological Engineers International (ASABE).

Interpretive Summary: none

Technical Abstract: Spectral signatures of Salmonella serotypes namely Salmonella Typhimurium, Salmonella Enteritidis, Salmonella Infantis, Salmonella Heidelberg and Salmonella Kentucky were collected using Fourier transform infrared spectroscopy (FT-IR). About 5-10 µL of Salmonella suspensions with concentrations of 108 cfu/mL were placed in direct contact with a diamond attenuated total reflection (ATR) crystal. Spectra were recorded from 4000 cm-1 to 525 cm-1 wavenumber (wavelength?) with the resolution of 4 cm-1 and data spacing of 1.928 cm-1. Collected spectra were subtracted from the background spectra of the empty diamond crystal surface. Principal Component Analysis (PCA) was conducted individually at five different spectral frequency regions (800- 525cm-1, 1200-800 cm-1, 1800-1200 cm-1, 2200-1800 cm-1 and 3000-4000 cm-1) to differentiate between the Salmonella serotypes based differences in the spectral features of biomolecular structural macromolecules of the bacterial isolates. PCA was also used to define the natural clusters in the data sets and to describe the difference between the sample clusters. At the 1800 – 1200 cm-1 region, PC1 distinguished 95% and PC2 distinguished 3 % of the serotypes, resulting in a maximum classification of 98 %. For all the five Salmonella serotypes, frequencies between 1000-1150 cm-1 and 1170 -1280 cm-1 had higher loading values which showed the significant contribution of these frequencies in the serotype classification. Findings indicate that PCA might serve as a useful tool for disinguishing between Salmonella seroypes.