Rapid evaluation method of eating quality based on near-infrared spectroscopy for composition and physicochemical properties analysis of rice grains

The eating quality of rice is a trait of high interest to breeders and consumers, but the existing detection methods are damaging, time consuming, or subjective. The composition and physicochemical properties in breeding are closely related to the eating quality, and near-infrared spectroscopy (NIRS) is a rapid and nondestructive detection method. In this study, we established NIRS models for the composition (amylose, protein, and fat content) and physicochemical properties (gel consistency and alkali spreading value) of rice grains. The model determination coefficients were > 0.86 and were analyzed for 112 japonica rice. The results showed that the composition and physicochemical properties were highly correlated with the eating quality of japonica rice. High eating quality with soft and sticky cooked rice had lower amylose content, lower protein content, higher fat content, higher gel consistency, and higher alkali-spreading value. A rapid and nondestructive evaluation method for eating quality using rice grains was developed with related composition and physicochemical properties, and the determination coefficient of predicted eating quality and traditional sensory evaluation score was 0.8741. Thus, the proposed method enables effective screening of desired eating quality according to different compositions and physicochemical, and eating qualities based on NIRS without preprocessing rice grains (with husk) in early generation breeding, which can shorten the breeding period.