The Establishment Of Near-infrared Model Of Peanut Fatty Acids And The Screening Of High Oleic Acid Germplasm

Peanut is one of the most important economic crops in China.With the continuous development of peanut industry in our country,the requirement of peanut is increasing and the demand is also increasing.In recent years,the peanut varieties with high oleic acid content have been occuping the main peanut market because of many advantages like excellent antioxidant stability,as well as high nutritional value and economic value.Therefore,peanut breeding of high oleic acid has become one of the important tasks of peanut breeding work.Looking for a way to speed up the breeding process,shorten the breeding cycle,improve the efficiency of peanut breeding techniques,and reduce the burden of work of scientific research personnel,has become the first problem in the current peanut breeding work.This means that the cooperation between efficient,fast and accurate near infrared spectroscopy analysis technology and peanut breeding will become a kind of trend in development in the future.This study used a randomized block design with 336 peanut germplasm materials,then,took a measurement for 22 main agronomic traits,including flowering period,growth period,plant height,branch length,branch number,weight of 100 pods,weight of 100 kernels,and the productivity of one plant,ect,and a total of 218 outstanding peanut germplasm resources were selected to screen.Then,we determined 3 kinds of fatty acids content in the peanut samples by gas chromatography technique,and treated the 218 peanut samples by using 3 different regression techniques and 3 different scattering techniques.Finally,3 near infrared spectroscopy models were established respecively for the determination of peanut oleic acid,linoleic acid and palmitic acid.We could screen the high oleic acid peanut germplasm using this three models after validating.The main results of this study are as follows:1)Near-infrared spectroscopy model testing peanut oleic acid content: the standard error of calibration?SEC?was 0.660,the standard error of validation?SECV?was 0.667,the cross-validation correlation coefficient?1-VR?was 0.842,and the egression squared?RSQ?was 0.829.According to the results of the validation of the model and the determination results of unknown peanut samples,it can be sure that the model can be used to determine the content of oleic acid in peanut.2)Near-infrared spectroscopy model testing peanut linoleic acid content: the standard error of calibration?SEC?was 0.565,the standard error of validation?SECV?was 0.588,the cross-validation correlation coefficient?1-VR?was 0.886,and the egression squared?RSQ?was 0.857.According to the results of the validation of the model and the determination results of unknown peanut samples,it can be sure that the model can be used to determine the content of linoleic acid in peanut.3)Near-infrared spectroscopy model testing peanut palmitic acid content: the standard error of calibration?SEC?was 0.890,the standard error of validation?SECV?was 0.883,the cross-validation correlation coefficient?1-VR?was 0.819,and the egression squared?RSQ?was 0.853.According to the results of the validation of the model and the determination results of unknown peanut samples,it can be sure that the model can be used to determine the content of palmitic acid in peanut.4)In the process of the establishment of 3 near infrared spectrum analysis models of peanut oleic acid,linoleic acid and palmitic acid,our study used the Improved Partial Least Squares regression?MPLS?.Compared with Partial Least Squares regression?PLS?and Principal Component Regression analysis technique?PCR?,the Improved Partial Least Squares regression is more suitable for building the near infrared spectroscopy model of peanut fatty acid.In the scattering processing,this research adopted the technique of the Standard Normal Variant and Detrend Only?SNV+Detrend?.The result is more ideal gotten by using this way than using Standard Normal Variant?SNV?and NONE in Scatter part.5)Validating the three models of oleic acid,linoleic acid and palmitic acid in peanut with unknown peanut samples,and the maximum absolute error of predicted value and chemical value were 1.33%,1.21% and 1.43%,respectively,no more than 1.5%.It is further proved that these 3 models can be used for the rapid determination of fatty acids in peanut.6)The content of oleic acid in 200 peanut germplasm was determined by the near infrared spectrum analysis model,and a total of 13 high oleic acid peanut germplasms with oleic acid content of more than 70% were screened:FB₄?12L₁₇?12L₂₀?12L₂₃?12L₃₃?12L₃₄?12L₃₆?12L₃₇?12L₄₉?12L₅₀?12L₅₂?12L₅₃?12L₅₄.