Classification of broiler breast fillets according to storage and to freeze-thaw treatment using near infrared spectroscopy and multivariate analysis

Authors: SOHN, M - UGA; Zhuang, Hong; Himmelsbach, David; Windham, William

Submitted to: Annual Meeting of the Institute of Food Technologists
Publication Type: Abstract Only
Publication Acceptance Date: 12/10/2008
Publication Date: 6/6/2009
Citation: Sohn, M., Zhuang, H., Himmelsbach, D.S., Windham, W.R. 2009. Classification of broiler breast fillets according to storage and to feeze-thaw treatment using near infrared spectroscopy and multivariate analysis. Annual Meeting of the Institute of Food Technologists.

Interpretive Summary: Visible/near-infrared (NIR) spectroscopy has shown potential for successfully classifying broiler breast fillets according to their texture properties. Freshness and shelf life are also important quality characteristics of boneless skinless chicken breast products in the marketplace. This study deals with the classification of chicken breast meats according to their storage time (shelf life) and storage method using NIR spectroscopy and multivariate analysis. Broiler fillets deboned 2hr postmortem were either stored in a refrigerator (3oC) for 14 days or in a freezer (-20oC) for 6 days and then in a refrigerator for 24 hours (freeze-thaw treatment). The refrigerated breasts were sampled after 0, 2, 7, 14 day storage, and the freeze-thaw breasts sampled after 7 days. NIR spectra of the samples were collected in reflection mode over a range of 400-2500 nm in 0.5 nm intervals. Spectral data were preprocessed using a second derivative followed by multiplicative scatter correction and mean centering. Principal component analysis (PCA) was performed for the classification. PCA results showed a good separation between individual storage groups. The first two PCs explained about 85% of x-variables. High loadings for the classification were observed at 440, 560 nm due to discoloration, at 1414, 1900 nm due to water, and at 1386, 1724, 1770 nm due to fat and protein. This indicated changes in components as well as in color during storage. The fresh meats and the freeze-thaw meats were clearly distinguished by two PCs, explaining over 90% of x-variables. The most highly correlated wavelengths were 438, 560 nm for PC1 and 1412, 1904 nm for PC2. The results suggested that NIR spectroscopy combined with multivariate analysis could be used for rapid classification of broiler breast fillets according to their storage time and to freeze-thaw treatment.

Technical Abstract: Visible/near-infrared (NIR) spectroscopy has shown potential for successfully classifying broiler breast fillets according to their texture properties. Freshness and shelf life are also important quality characteristics of boneless skinless chicken breast products in the marketplace. This study deals with the classification of chicken breast meats according to their storage time (shelf life) and storage method using NIR spectroscopy and multivariate analysis. Broiler fillets deboned 2hr postmortem were either stored in a refrigerator (3oC) for 14 days or in a freezer (-20oC) for 6 days and then in a refrigerator for 24 hours (freeze-thaw treatment). The refrigerated breasts were sampled after 0, 2, 7, 14 day storage, and the freeze-thaw breasts sampled after 7 days. NIR spectra of the samples were collected in reflection mode over a range of 400-2500 nm in 0.5 nm intervals. Spectral data were preprocessed using a second derivative followed by multiplicative scatter correction and mean centering. Principal component analysis (PCA) was performed for the classification. PCA results showed a good separation between individual storage groups. The first two PCs explained about 85% of x-variables. High loadings for the classification were observed at 440, 560 nm due to discoloration, at 1414, 1900 nm due to water, and at 1386, 1724, 1770 nm due to fat and protein. This indicated changes in components as well as in color during storage. The fresh meats and the freeze-thaw meats were clearly distinguished by two PCs, explaining over 90% of x-variables. The most highly correlated wavelengths were 438, 560 nm for PC1 and 1412, 1904 nm for PC2. The results suggested that NIR spectroscopy combined with multivariate analysis could be used for rapid classification of broiler breast fillets according to their storage time and to freeze-thaw treatment.