Kjeldahl method used in many materials from various plant parts to determine protein contents, is laborious and time-consuming and utilizes hazardous chemicals. Near-infrared (NIR) reflectance spectroscopy, a rapid and environmentally benign technique, was investigated as a potential method for the prediction of protein content. Near-infrared reflectance spectra(1100-2400 nm) of coarse cereal grains(n=100 for each germplasm) were obtained using a dispersive spectrometer as both of grain itself and flour ground, and total protein contents determined according to Kjeldahl method. Using multivariate analysis, a modified partial least-squares model was developed for prediction of protein contents. The model had a multiple coefficient of determination of 0.99, 0.99, 0.99, 0.96 and 0.99 for foxtail millet, sorghum, millet, adzuki bean and mung bean germplasm,respectively. The model was tested with independent validation samples (n=10 for each germplasm). All samples were predicted with the coefficient of determination of 0.99, 0.99, 0.99, 0.91 and 0.99 for foxtail millet, sorghum,millet, adzuki bean and mung bean germplasm, respectively. The results indicate that NIR reflectance spectroscopy is an accurate and efficient tool for determining protein content of diverse coarse cereal germplasm for nutrition labeling of nutritional value. On the other hands appropriate condition of cereal material to predict protein using NIR was flour condition of grains.
MATERIALS AND METHODS
RESULTS AND DISCUSSION
ACKNOWLEDGEMENT
REFERENCES