Study On Screening Fingerprinting Information And Discriminant Of Origin Traceability Of Fuji Apple In China

In recent years,as people pay more attention to the geographical origin of agricultural products,the geographical origin traceability of agricultural products has attracted extensive attention of scholars at home and abroad.As one of the agricultural products,apple is widely planted around the world,and China is the largest apple producing country in the world.In this study,Fuji Apple was selected as the experimental subjects.In 2017 and 2018,a total of 164 Fuji apple samples were collected from three main apple producing regions in China,the Bohai Bay（BHB）,Loess Plateau（LP）and Northwest（NW）.The organic compounds,stable isotope ratios,multi elements,volatile compounds,physical and chemical index in the samples were determined by liquid chromatography,ion chromatography,isotope ratio mass spectrometry,inductively coupled plasma mass spectrometry,gas chromatography-mass spectrometer,automatic potentiometric titrator and other instruments.The contents characteristics of indicators,the correlation between indicators and regional conditions,the regional differences of the indicators,and the effects of various factors（year,regional and their interaction）on the variation of the content of organic compounds,stable isotope ratios and multi elements were systematically analyzed.After that,the indicators which were closely related to the regions were selected and combined with chemometrics methods to classify the origin of training set samples and establish the discriminant model of the samples from the three producing regions,moreover,the stability and effectiveness of the discriminant models were verified by the prediction set（unknown sample）.So as to explore the effectiveness of fingerprinting information of organic compounds,stable isotope ratios,multi elements,volatile compounds,physical and chemical index in the geographical origin traceability of Fuji apple.The main conclusions were as follows:1.Sorbitol,glucose,fructose and sucrose were closely related to the regions.By the effective indicators,sorbitol,glucose and fructose,the correct classification rate of training set of BHB-NW sample group was 87.1%and the correct recognition rate of the discriminant models on prediction set was 82.6%.In addition,it was found that the organic acids were not suitable for distinguishing the origin of the samples.2.Theδ¹³C value,which was closely related to the regions,was used to establish the discriminant models of samples from different regions.By the discriminant models,the correct recognition rate of prediction set of BHB-LP samples group was 92.3%and the correct recognition rate of prediction set of BHB-NW samples group was 91.3%.In addition,the correct recognition rate of the discriminant models on BHB samples of prediction set in different samples groups（BHB-LP,BHB-NW and BHB-LP-NW）was 100%.3.Combined with linear discriminant analysis,the element of Mg,K,Mn,Fe,Co,Cu,As,Sb,Mo,Pr,Nd,Eu,Gd,Dy and Ho,which closely related to the regions,had a correct classification rate of 85.9%-97.1%for the training set samples grouped by two regions（BHB-LP,LP-NW and BHB-NW）,and the correct recognition rate of the discriminant models on prediction set samples was 84.8%-95.7%.Although the correct classification rate of the training set samples from the three regions（BHB-LP-NW）was only81.3%,the correct recognition rate of the discriminant models on the prediction set samples of the three regions can reached to 92.7%.4.Through the combination of three fingerprinting information（organic compound,stable isotope,and multi element）,the correct classification rate of training set from the three regions（BHB-LP-NW）was 89.4%,and the correct recognition rate of the discriminant models on prediction set was 87.8%,moreover,the effective indexes selected by discriminant analysis were glucose,fructose,sucrose,K,Mn,Co,Mo and Sb.5.The indicators,which closely related to the regions（fruit shape index,core size,b*,h^o,propyl propionate,hexyl hexanoate and n-pentadecane）,combined with linear discriminant analysis,had a correct discrimination rate of 87.8%-94.1%for the two regions sample group（BHB-LP,LP-NW and BHB-NW）,which showed that it was feasible to use the volatile compounds and physical and chemical index as the indicators for identifying the origin of Fuji apple.Therefore,the fingerprinting information has its own characteristics and the samples from the three regions（BHB,LP and NW）can be stably and effectively classified and predicted by the fusion of the three kinds of fingerprinting information（organic compounds,stable isotope ratios and multi elements）.The results of this research can provide reference for the origin traceability research of apple.