Study On The Association Between Physicochemical Properties Of Individual Rice Grains And Eating Quality Based On Near-infrared Spectroscopy

Rice is the staple food for nearly 50%of the global population,providing a large variety of nutrients,such as starch,protein,and fat.The taste of rice is one of the most direct indicators for consumers to judge the merits of milled rice.Eating quality is a comprehensive quality influenced by various physicochemical indicators in the processing,appearance,and nutrition of rice,as well as subjective factors of consumers,so it has always been the focus and difficulty of rice quality breeding.Therefore,it is important to develop an intelligent method for the identification of physical and chemical characteristics of rice quality and eating quality to improve breeding efficiency.The near-infrared spectroscopy(NIRS)technology,as a rapid and nondestructive analytical technique,is widely used for analyzing the physicochemical properties of rice.The taste meter,which represents the eating quality,also realizes the quantitative analysis by combining the NIRS and sensory score of rice.However,the application of these methods is limited by the materials and cannot analyze the physicochemical characteristics of individual rice grains and the eating quality,which restricts the analysis of the early genetic behavior of quality traits and its application in breeding.In this study,based on the self-developed NIRS intelligent sensing platform QSC01 for single-grain rice quality traits,the "standard grain" used for stability and repeatability evaluation of testing instruments were produced using 3D printing technology as the "internal standard" for self-test of the platform system.The physicochemical characteristic NIRS models were established that were suitable for the nondestructive dynamic testing and sorting of individual rice grains,and the correlation between rice physicochemical indicators and food quality was studied.The effect of the separation of single-grain physicochemical characteristics of hybrid rice on the eating quality was studied,which provided a phenotype identification technology for the genetic breeding of "high nutrition and good taste" rice varieties.The main findings are as follows:1.Completion of the stability self-test system of QSC-01 and establishment of NIRS models for single-grain rice physicochemical analysis.(1)Based on 3D printing technology,"standard grains" with similar shape,stable properties,uniform composition,and a concentration gradient distribution were constructed.The "standard grains" can quickly evaluate the stability of the instrument operation to ensure the reliability of analysis data.(2)Establishment of models for single-grain rice quality physicochemical indicators.The cross-test models of single-grain amylose content(AC),protein content(PC),fat content(FC),gel consistency(GC),alkali spreading value(ASV),grain length,length-width ratio,and brown rice rate were built with regard to the coefficients of determination R2≥0.8,which can satisfy the screening of the quality traits of individual rice grains.(3)The fusion of transflectance spectrum and difuse reflection spectrum,as well as the near-infrared spectrum and ultraviolet-visible spectrum.The results showed that when the absorbance of different spectrometers was close,the sample information can be complementary in the same and different bands respectively.Spectral fusion can further improve the accuracy of the quality trait analysis.The above research is crucial for the QSC-01 device to achieve accurate and stable analysis of multiple quality traits of individual rice grains.2.Nondestructive evaluation of rice eating quality based on NIRS.NIRS was used to detect the physicochemical properties(AC,PC,FC,GC,and ASV)of rice grains,and a correlation analysis was conducted based on the taste values detected by the taste meter.It was found that the higher the AC and PC,the lower the FC,ASV,and GC,and the worse the eating quality of conventional japonica rice and hybrid indica rice.The correlation between the physicochemical characteristics of conventional japonica rice grains and the eating quality model was good(R2=0.8345),and the predicted eating quality was in good agreement with the sensory scores(R2=0.8741).The ranges of AC,PC,FC,GC,and ASV of japonica rice with a high taste quality were 3.4-15.3%,7.39.1%,2.7-3.8%,65.3-97.4 mm,and 5.2-7.0,respectively.Compared with conventional rice,the eating quality model of hybrid indica rice was less satisfactory(R2=0.7282),and the correlation between predicted food quality and the sensory evaluation score was lower(R2=0.5404).3.Aiming at the unclear correlation between the physicochemical traits of hybrid indica rice and eating quality,the QSC-01 test found that the quality characteristics among individual rice grains of hybrid indica rice varieties were separated,and the higher the variation(degree of dispersion)in single-grain quality,the worse the eating quality.The results were also verified using chemical methods and static NIRS singlegrain quality detection.Moreover,the addition of the differences in single-grain composition improved the R2 of the eating quality model constructed using the composition from 0.657 to 0.850,which can be used for the nondestructive evaluation of the eating quality of hybrid indica rice.Among the parental lines of hybrid indica rice,taking the Wx allele of AC as an example,it was found that the Wxa/Wxb combination groups had a poorer eating quality than the Wxb/Wxb combination groups,and the sterile lines with a high AC had a negative effect on the eating quality of the hybrid combination.Meanwhile,the physicochemical characteristics of the sterile lines contributed 74.4%to the eating quality of hybrid indica rice,indicating that the selection of high-quality sterile lines is the key to improving the eating quality of hybrid indica rice.This study provides technical support for predicting the eating quality of hybrid indica rice using single-grain NIRS.