Main content area

An investigative study on differentiation of citrus fruit/fruit juices by UPLC-QToF MS and chemometrics

Jandrić, Z., Cannavan, A.
Food control 2017 v.72 pp. 173-180
chemometrics, data collection, discriminant analysis, fraud, grapefruits, ionization, least squares, mandarins, mass spectrometry, models, orange juice, oranges, principal component analysis, pummelos, rapid methods
In the study presented, the potential of coupling chemometrics and mass spectrometry (UPLC-QToF MS) data for distinguishing the origin and variety of citrus fruit/fruit juices was investigated. Principal component analysis (PCA), partial least squares discriminant analysis (PLS-DA) and soft independent modelling by class analogy (SIMCA) were employed. Clustering of citrus fruits (orange, grapefruit, mandarin, and pomelo) and oranges of various geographical/botanical origin was revealed using PCA on data produced in both negative (goodness of fit, R2(cum) = 0.67–0.92, and predictability, Q2(cum) = 0.56–0.88) and positive (R2(cum) = 0.58–0.85 and Q2(cum) = 0.47–0.80) ionisation mode. PLS-DA and SIMCA confirmed the results (with 100% recognition ability obtained for citrus fruits/orange models and fresh squeezed commercial orange juice, while classification success of 80% was achieved for commercial orange juice prepared from concentrate) and showed that the category models for the class can be sensitive and highly specific. Characteristic compounds responsible for the discrimination were identified. The applicability of the models was tested with an external data set of fruit juices adulterated with other fruit juices down to 1% and diluted with water down to 5%. Using Coomans' plots, adulterated samples were easily distinguished from authentic samples showing the possibility of applying this method as a rapid screening technique to trace or confirm the origin of citrus fruit/fruit juices and detect fraud.