Authentication of juices from antioxidant and chemical perspectives: A feasibility quality control study using chemometrics

In this work, Brazilian juices (n = 38) from distinct botanical species were analyzed for the physicochemical properties, major phenolic classes, and antioxidant activity using high-throughput assays. Principal component analysis [PCA] was applied to study the data structure, while classification methods based on partial least squares-discriminant analysis [PLS-DA] and dual data-driven PCA/soft independent modeling of class analogy [DD-SIMCA] were used to predict the class membership of juices. In addition, multiple linear regression [MLR] models were proposed to explain the antioxidant activity of juices. PLS-DA was successfully used to authenticate the class membership of juices, enabling the identification of the main variables responsible for the discrimination. Similarly, DD-SIMCA was shown to be useful for the authentication of juices. Additionally, the main phenolic classes responsible for each of the antioxidant activity were revealed by MLR. Overall, the characterization of juices was reached by the application of relatively simple analytical methods supplemented with modern chemometric tools.