Submitted to: Near Infrared Spectroscopy International Conference Proceedings
Publication Type: Proceedings
Publication Acceptance Date: 11/7/2009
Publication Date: 11/7/2009
Citation: Sohn, M., Zhuang, H., Lawrence, K.C., De Haseth, J.A. 2009. NIR spectroscopy for determining soy contents in processed meat products. Near Infrared Spectroscopy International Conference Proceedings. November 7-12, Bangkok, Thailand. p. 235-239.
Interpretive Summary: Soy products such as soy concentrate, soy protein and soy grits, are commonly added to meat and poultry products to enhance meat functionality (texture, water and fat retention). However, soy products can also cause allergies, especially in infants and young children, and the allergies can be mild to life-threatening. The addition of soybean to meat products is regulated. In order to ensure that a meat product is within the legal limitations and prevent potential frauds, different analytical methods for the determination of soybean proteins in meat products have been developed. These methods are based on electrophoretic, immunochemical, reaction/extraction, or chromatographic techniques, and require tenuous sample preparation, and time consuming. This study was conducted to develop a rapid, non-destructive NIR spectroscopic method to predict soy contents in meat product. Our results show that there is close relationship between soy contents in meat and the changes in NIR spectra, and that NIR spectroscopy could be used to predict soy contents rapidly and non-destructively in further-processed meat products.
Technical Abstract: Soy products such as soy concentrate, soy protein and soy grits are used as a meat extender in processed meat products to improve meat texture. However, soy allergies are one of the common food allergies, especially in infants and young children, and can be mild to life-threatening. The United States Department of Agricultural (USDA)-Food Safety Inspection Service (FSIS) has been routinely requested to determine soy contents in further-processed meat products, such as ground meat, meat patties and frankfurters. Currently analytical method used is time consuming (about 4 hr) and lacks in accuracy. This study was conducted to develop a rapid NIR spectroscopic method to predict soy contents in meat product. Ground beef (93%-87% lean meat) and four different soy products were used for the study. The ground meat was mixed with the soy products, providing a range of soy contents from 0 to 30%. A total of 70 beef-soy mixture samples were prepared. Samples were packed in a round cell and scanned on a NIRSystems 6500 (FOSS NIRSystems, Inc., Laurel, MD) over a range of 1100 to 2498 nm at 2 nm intervals. Partial least squares regression (PLSR) was used for model development. The best preprocessing of the data for the model was Savitzky-Golay 1st derivative (3rd polynomial and 15 points) followed by multiplicative scatter correction and mean centering. The PLS regression resulted in an R2 of 0.99, RMSEC of 0.83% and RMSECV of 0.93% using four latent variables. The most highly correlated wavelengths for the model were 1694nm, 2055, 2166 and 2208 nm, which are assigned to protein bands. The results demonstrated that NIR spectroscopy could be used to predict soy contents in further-processed meat products.