Prediction of rice starch quality parameters by near-infrared reflectance spectroscopy

A rapid predictive method based on near-infrared spectroscopy (NIR), was de veloped to measure rice starch quality parameters. A calibration set of 100 samples and validation set of 62 samples of rice flour of Chinese genotype s was used. Results of partial least squares modeling indicated that NIR wa s reasonably accurate in predicting apparent amylose content (AAC) (standar d error of prediction, [SEP] = 1.39 percentage units, coefficient of determ ination [R-2] = 0.91); pasting parameters of setback (SB) (SEP 13.6 RVU, R- 2 = 0.92), and breakdown (BD) (SEP = 10.2 RVU, R-2 = 0.88); and gelatinizat ion peak temperature (T-p) (SEP = 1.33 degreesC, R-2 = 0.89). Gel consisten cy (GC), cool paste viscosity (CPV), gelatinization onset temperature (T-o) , and textural properties of chewiness, hardness and gumminess, were modele d less well with R-2 between 0.75 and 0.86. NIR analysis is sufficiently ac curate for routine screening of large numbers of samples in early generatio n selection in rice breeding programs.