Study On Near-infrared Reflectance Spectroscopy Model Of Measuring Amylose Content In Rice

Rice is an indispensablefood crop in China.The quality of cooking and eating is the most important index for evaluating rice quality.Amylose content?AC?is the most important indicator that greatly affects the cooking and eating quality of rice.The traditional method to measure amylose content in rice is complex,time-consuming and destructive,and it is difficult to measure lots of rice samples quickly and efficientlyin real time.Near-infraredreflectance spectroscopy?NIRS?is a promising technique with accurate,fast and nondestructive analytical potentials,which can measure the content of different organic compounds.Recently,the technique has gained popularity as an efficient screening tool for AC analysis in rice.The purpose of this study is to build up and select the available models which can measure the AC value rapidly and accurately in rice,based on the abundant rice resources and the use of the DA7200 Diode ArrayAnalyzer and the Unscrambler software.With the influences of different conditions on the accuracy of the models?including the types of samples,the range of AC values,preprocessing spectral methods and the proportion of eliminating outliers?,we explored and summarized the technical points of NIRS and selected the available models for rice breeding.The main conclusions are as follows:1 Selection of available modelsIn this study,we have built up 120 NIRS models of three sample forms.The milled rice flours model is found to be the most suitable model to measure amylose content in rice,by comparing different evaluation index of models,the accuracy,the required weight of unknown samples,and the difficulty of sample preparation.?1?We selected 14 available milled rice flours models.The detection method for unknown samples by these models is as follows:Firstly,prepare the milled rice flours samples of 7-14 g.Secondly,put the samples at 42°C more than 2 hours,a nd then place them at room temperature for more than 12 hours to make the moisture content of samples stable.Measure the milled rice flours samples through the 14 NIRS models simultaneously?20 models can be used for each measurement?,and export the results.The average of the14 measured values is the AC value.?2?If the unknown sampleswere sufficient,the brown rice models and the milled rice models would be availble.This study selected 3 available brown rice models and 4 milled rice models,and the detection method for each unknown sample by these models is as follows:Prepare the brown rice samples of 105-210 g?or milled rice samples of 120-240 g?.Measure the samples through the 3 brown rice NIR models?or 4 milled rice NIR models?simultaneously,and export the results.The average of the measured values is the AC value.2 Use of models in rice breedingIn this study,the sample types and SEP values of different models were vicissitudinous.The measured AC value of different materials in rice breeding would have different applicable models and different selection indicators.?1?For the selection of M2 populationof mutagenic rice materials,we recommended to use milled rice flours models because the selection target isa single plant and the number of rice is small.The mutagenic materialswill be valuable and selected when their AC value aresignificantly different from those of CK.That is,the measured AC value of the unknown samples should be less than or more than the CK value of 3.84 or more.?2?For the selection of F₂ generation of rice breeding materials,westillrecommended to use milled rice flours models because the selection target is also a single plant.The F₂generationmaterials are valuable and selected when their AC valueareabout 20%in South China.In addition,the AC value of early-seasonricecultivationis ususlly less 1-2%than that of late-season rice.That is,the unknown F₂ generationsamples can be identified as valuable materials when their measured AC value by NIRS models is 16.16-23.84%for late-season rice and 15.16-22.84%for early-season rice.The specific adjustment range in early and late season rice can depend onyour practical experience.?3?For the high-generation rice lines which are sufficient,three kinds of samples models?includingmilled rice flours,milled rice,brown rice?can be used.The high-generationmaterials are valuable and selected when their AC valueareabout 20%.For the late-season rice,the unknown samples can be identified as high-quality materials when their measured AC value by NIRS models is 16.16-23.84%for milled flours rice,16.12-23.88%for brown rice and 16.62-23.38%for milled rice.The rapid detection modelsbased on NIRSto measure amylose content of rice were constructedfirstly in South China.The results of the study have theoretical and practical significance for improving the early selection efficiency of rice amylose content traits and promoting high quality rice breeding.