Skip to main content
ARS Home » Research » Publications at this Location » Publication #183090
Title: PREDICTION OF MONOUNSATURATED, POLYUNSATURATED, AND SATURATED FATS BY NIR AND FT-NIR SPECTROSCOPY IN PROCESSED CEREAL PRODUCTS

Author
item Kays, Sandra
item VINES, LAURA - UNIVERSITY OF GEORGIA
item Kim, Yookyung
item KOEHLER, PHILIP - UNIVERSITY OF GEORGIA

Submitted to: Federation of Analytical Chemistry and Spectroscopy Societies Final Program
Publication Type: Abstract Only
Publication Acceptance Date: 6/29/2005
Publication Date: 10/10/2005
Citation: Kays, S.E., Vines, L.L., Kim, Y., Koehler, P.E. 2005. Prediction of monounsaturated, polyunsaturated, and saturated fats by nir and ft-nir spectroscopy in processed cereal products [abstract]. Federation of Analytical Chemistry and Spectroscopy Societies Final Program. P. 188

Interpretive Summary:

Technical Abstract: U.S. nutrition labeling legislation requires the declaration of total and saturated fat content on a product’s nutrition label, while declaration of mono- and polyunsaturated fat is voluntary. The accepted method for analysis of these components is extremely labor-intensive and time consuming. Previous work has shown that total fat in diverse processed cereal food products can be predicted accurately by near-infrared (NIR) reflectance spectroscopy; therefore, NIR reflectance spectroscopy was investigated as a rapid method for the analysis of saturated, polyunsaturated and monounsaturated fat. Fatty acid composition was determined in ground products by AOAC Method 996.01 and NIR spectra obtained using a dispersive NIR instrument (range 400-2498 nm) and a Fourier transform NIR instrument (range 1000-2500 nm). Individual chemometric models (n=72-73) for the prediction of saturated, polyunsaturated and monounsaturated fatty acids were developed for each instrument type with multivariate analysis software. Using the dispersive instrument, the standard error of cross validation and multiple coefficient of determination for the prediction of saturated fat were 1.08 (range 0.1-14.7) percent and 0.88, respectively; for polyunsaturated fat were 0.57 (range 0.2-22.0) percent and 0.98, respectively; and for monounsaturated fat were 1.5 (0.1-24.0) percent and 0.94, respectively. Performance of the models developed using the different instruments will be compared. All models were tested with an independent validation set of processed cereal products (n=34-36) and 90-100 percent of the validation samples were predicted within the accuracy required by U.S.A. nutrition labeling regulations. Thus, near-infrared spectroscopy has excellent potential for prediction of lipid classes in processed cereal products.