Calibration Models Of Near Infraredreflectance (NIR) For Rice Quality Characteristics And Genetic Association Analysis

Rice quality and genetic features are very important for the rice breeding. Rice quality generally divided into milling quality, appearance quality, eating and cooking quality and nutrition quality. Nowdays eating and cooking quality and nutrition quality get more and more attention. There are generally three evaluation index (amylose content, gelatinization temperature and gel consistency) for Rice cooking and eating quality; Nutritional quality indicators are protein content, amino acid compositions. To measure these qualities need many steps, high cost, time-consuming, especially for small sample, it is difficult to keep the original sample to grow again;It is urgent to use the lossless near infrared analytical technique for measuring the quality,At the same time understanding the rice quality of the genetic characteristics is important for the quality improvement of breeding. Our study use 797 rice germplasm resources, among them there are 275 fragrant rice varieties. Using near infrared grain analyzer to spectrum scan entry, and optimize the data of spectral band and statistical regression, finally determine the rice cooking quality and nutritional quality the best calibration model. Base on the near infrared quality model, we measure cooking and nutrition quality of 275 rice germplasm resources, meanwhile we also identify the genotype of 275 rice germplasm resources by using 147 SSR markers, and have analysis the correlation between amylose content, gelatinization temperature, gel consistency,protein content, amino acid content and SSR markers.Main result:1)rice quality characteristics model:the results indicate that rice flour protein content and amylose content model had better result, coefficient of determination of the calibration model were 0.946 and 0.974,1-VR were 0.951 and 0.963, standard error of calibration are respectively 0.303 and 1.338; Its external inspection work standard error (standard error of the performance, SEP) were 0.346 and 1.300 respectively, RPD> 4.0. RPDS (corresponding residual predictive deviation) is the ratio of standard deviation and work standard error RPDS (SD/SEP), this index is determine calibration model. Generally greater than 4.0 indicated the calibration model is very good, and the two models of scaling determination coefficient (RSQ) and cross validation decision coefficient (1-VR) are more than 0.950, shows that the performance of model was good, can be used for rapid determination of rice quality.2)The rice quality characteristics of SSR correlation analysis:57 genetic loci associated with rice quality traits. There are two SSR makers associated with protein content, five SSR makers associated with amylose content, six SSR makers related to gel consistency and three SSR makers associated with alkali value elimination, the rest 41 SSR makers were associated with total amino acids and 17 kinds of amino acids tag 41.There are 2 mark, RM129 and RM5807 related to gel consistency and protein, amino acid gene loci respectively on chromosome 1,2. There are four SSR makers at the same time were associated with two quality traits of gene loci. Such as on chromosome 6 SSR makers RM6467 was associated with gel consistency, amino acid gene loci; on chromosome 9 SSR makers RM23916 associated with amylose, gel consistency; on chromosome 11 SSR makers RM7120 associated with amylose, gel consistency gene loci. SSR makers RM27154 associated with proteins content and amylose content.These SSR makers provide a reference for rice quality breeding.