As a high-value tropical seasonal fruit,the origin information of durian(Durio zibethinus Murr.)has attracted the attention of importers and consumers in recent years.The establishment of the origin traceability of durian is of great importance for brand protection and the protection of informed rights of consumers.This research majored in durian,which collected from Malaysia,Thailand,Vietnam and Cambodia.Multivariate stable isotope analysis,mineral elements and Raman spectra combined with chemometrics was used to investigate and discriminate durian samples from four origins.ANOVA,two-way ANOVA,principal component analysis,linear discriminant analysis and artificial neural network were used as exploratory techniques and classification parametrics.The purpose of this study is to provide a technical solution for durian traceability and serve as a template for further research on the traceability of other agricultural products,especially plant-derived products.The main findings are as follows:The stable isotope and element content analysis combined with ANOVA were used for exploring the influence of maturity and varieties of durian on the measured data.Durian core and pulp using five stable isotopes(δ15N,δ13C,δ34S,δ2H andδ18O)and elemental contents:nitrogen(N%),carbon(C%)and sulfur(S%)were analyzed.The results showed that there were no significant difference in indexes exceptδ2H values in durian core andδ18O values in durian pulp.δ18O values in durian core andδ34S values and N%in pulp had significant differences among different varieties of Thai durian,and there were significant differences inδ34S values in core,δ34S values and C%in pulp in durians from Malaysia with different varieties.Durian core and pulp using stable isotopes and elemental contents between four origins were analyzed.It turned out indexes had significant difference expectδ18O values in durian core,δ13C values and S%in pulp.LDA achieved durian origin accuracy rates of 98.6%and 97.3%for the testing and training sets respectively by combining stable isotopes and elemental contents.ANN gave higher accuracy rates than LDA,correctly identifying durian origin with an accuracy of 100%for the training set,and 94.4%on the testing set.Durian core and pulp using Inductively Coupled Plasma Mass Spectrometer(ICP-MS)for 28 kinds of mineral elements content were analyzed,and substituted the mineral elements with significant differences in durian into the LDA and ANN.The results showed that the accuracy of testing and training sets based on LDA is 91.8%and 90.4%,and the accuracy of ANN is 96.1%and 95.5%.The Raman spectra of durians from fours origins were measured,characteristic peaks at 502 cm-1、642 cm-1、910 cm-1、1052 cm-1、1080 cm-1、1206 cm-1、1459cm-1 were selected and substituted into LDA.The results showed the accuracy of both the testing sets and training sets were 80%.Coupling analysis of indicators with significant differences in stable isotope values,element content and mineral element content of durian core and pulp,The results showed that the accuracy of LDA and ANN were 100%,which indicated that the coupling of multiple technologies could distinguish the identification of durian origin. |