Descripción
|
|
---|---|
Recent work has shown that chemical composition and nutritional value of soybean meals (SBM) differ among origins. Near infrared reflectance spectroscopy (NIRS) has demonstrate its potential to identify and classify food products, and colour has also been used to detect improper processing conditions. The objective of this work was to evaluate NIRS and colour methods and its application in the feed mill industry in order to permit rapid and accurate origin (Argentine, ARG; Brazil, BRA; USA) identification of commercial SBM. A total of 290 SBM unground samples of known origin (ARG, n=84; BRA, n=74; USA, n=132) were scanned using a spectrophotometer (Foss NIRSystems 5000) over the range 1100-2500 nm. Colour (CIELAB system) was recorded using a Minolta colorimeter (model CR-300). On the basis of spectral data, a Principal Component Analysis was performed and outliers with a Mahalanobis distance greater than 3 were removed. The three sets of SBM spectra were split using the CENTER algorithm (WinISI v.1.50) into a calibration set (ARG, n= 61; BRA, n=53; USA, n=86) used to develop linear discriminant models (PLS2-DA procedure), and an external validation set (ARG, n=20; BRA, n=16; USA, n= 43). The best discriminant model yielded a SECV of 0.2515 and over 97% of correctly classified samples in the training set (100% for ARG; 96% for BRA, and 98% for USA, respectively), whereas in the validation step 100% of the samples were correctly classified. Colour parameters of SBM samples differed (P<0.001) among origins, with the BRA meals being redder than the ARG and USA meals (7.6 vs. 6.5 vs 5.9, respectively).The yellowness and luminosity were higher for the USA meals (71.1 and 26.5) than for the South American meals (67.3 and 25.1 for ARG; 65.3 and 24.4 for BRA), respectively. The results obtained show that NIRS is a reliable method of origin identification of commercial SBM samples at the reception stage in the feed mill industry, and that colour can help to confirm these results. | |
Internacional
|
Si |
Nombre congreso
|
99th Poultry Science Association Annual Meeting |
Tipo de participación
|
960 |
Lugar del congreso
|
Denver |
Revisores
|
Si |
ISBN o ISSN
|
08-134-2929-3 |
DOI
|
|
Fecha inicio congreso
|
11/07/2010 |
Fecha fin congreso
|
15/07/2010 |
Desde la página
|
657 |
Hasta la página
|
657 |
Título de las actas
|
Poultry Science, 89 (1): 657 |