Metabolic Fingerprinting for Verifying Authenticity in Raw and Toasted PDO Hazelnuts

Wednesday, May 1, 2024

11:35 AM – 11:55 AM EDT

Room: 520c

Presenting Author(s)

Berta Torres-Cobos (she/her/hers)

Ms.; PhD student
University of Barcelona, INSA-UB, Catalonia, Spain

Co-Author(s)

AD

Antonio G. Di Corato

Mr.; student
University of Barcelona, Spain
CM

Carlo Massoli

Mr.; student
University of Barcelona, Spain
LL

Lorenzo Laterza

Mr.; student
University of Barcelona, Spain
SN

Soriana B. Nicotra

Ms.; student
University of Barcelona, INSA-UB, Spain
AR

Agustí Romero

Dr.; Researcher
Institute of Agrifood Research and Technology (IRTA), Spain
MR

Mercè Rovira

Dr.; Researcher
Institute of Agrifood Research and Technology (IRTA), Spain
FG

Francesc Guardiola

Dr.; Professor
University of Barcelona, INSA-UB
Santa Coloma de Gramenet (Barcelona), Catalonia, Spain
Alba Tres

Dr.; Researcher
University of Barcelona, INSA-UB
Santa Coloma de Gramenet (Barcelona), Catalonia, Spain

Co-Author(s)

SV

Stefania Vichi

Dr.; Professor
University of Barcelona, INSA-UB
Santa Coloma DE Gramenet, Catalonia, Spain

The considerable economic value of hazelnuts makes them susceptible to fraud, especially in counterfeiting their geographical and botanical origin. This involves mislabeling of either the declared cultivar or provenance. Given hazelnuts' widespread use in various forms, such as raw and semi-processed commodities utilized at an industrial level, fraud poses a significant concern for the industrial sector, particularly for products recognized under EU designations like Protected Designation of Origin (PDO) and Protected Geographical Indication (PGI). Therefore, ensuring hazelnut authenticity is paramount for consumer protection.
Recent research highlights the potential of the hazelnut unsaponifiable fraction's composition, influenced by both genetic and environmental factors. Fingerprinting this fraction through gas chromatography-mass spectrometry and pattern recognition analysis is a promising tool for geographical and varietal authentication. However, these results have been obtained on raw samples produced under known management, thus validation is essential on commercially produced hazelnuts under standard agronomic management, including those subjected to processing like thermal treatment.
In this study, we present the outcomes achieved through the development of authentication models designed to distinguish hazelnuts from PDO "Avellana de Reus" (Spain) form commercial Turkish hazelnuts in both raw and roasted forms. Two binary partial least square-discriminant analysis (PLS-DA) models are developed for (i) PDO and Turkish raw hazelnut samples (n=60) and (ii) the same samples after roasting. For this, each sample set is randomly split into training (80%) and validation (20%) sets. The models developed with the training sets are externally validated by classifying the samples of the validation sets. External validation assesses the potential of unsaponifiable fraction fingerprinting for PDO hazelnut authentication and its ability to detect fraud in semi-processed hazelnut products used by the industry. The results of this prospective study will guide further testing on a larger sample set to establish robust classification models.