Characterization and classification of lulo (Solanum quitoense Lam.) fruits by ripening stage using partial least squares discriminant analysis (PLS-DA)

Abstract

Lulo or naranjilla (Solanum quitoense Lam.) is a tropical fruit with great potential for its contents of antioxidant and biofunctional compounds and sensory characteristics. Nowadays, the different methodologies to classify the ripening stage of lulo fruits are prone to bias and can hinder adequate characterization of the fruit maturity stage as they do not use measurements. The aim of this research was to define an accurate method for classifying lulo fruits by ripening stage based on non-destructive parameters and to determine their main characteristics according to the ripening stage. Hierarchical cluster analysis was carried out to classify fruits according to their maturity index (MI) into two (MI2) and three (MI3) homogeneous groups of individuals. Using partial least squares discriminant analysis (PLS-DA), with the non-destructive parameters showing significant differences between groups, classification functions by ripening stage were established. The PLS-DA correctly classified 89.47% of the fruits in the MI2 classification and 78.95% in the MI3 classification. The predictive power of the models was tested with fruits other than those used to establish the prediction equations, obtaining a correct classification in 75% of the cases. It is possible to classify lulo fruits objectively with a limited number of non-destructive parameters that constitutes a useful tool from harvesting to consumption.