Artificial Intelligence smelling by GC×GC-MS/FID as a tool for food aroma blueprinting: the unique aroma of Brazilian Olive Oils

Extra virgin olive oil (EVOO) has a complex aroma, which is related to the cultivar, the pedoclimatic conditions where trees are grown, olives maturation stage, and extraction process applied. Brazil is the second largest importer of olive oil in the world, but, despite the high consumption, the growing of olives in the country and, therefore, the extraction of oil, are quite recent and small in relation to the size of its domestic market. This work aimed to explore the Brazilian EVOO volatilome by GC×GC-MS/FID and evaluate the effectiveness of chromatographic blueprinting to recognize characteristic patterns of odorants for different cultivars (Arbequina and Koroneiki) and production regions (Rio Grande do Sul and Serra da Mantiqueira). Samples from Arbequina and Koroneiki cultivars from the two main production regions in Brazil in 2021 and 2022 harvest were analyzed. GC×GC was performed with a polar × semi-polar column combination followed by qMS/FID parallel detection. Untargeted/targeted fingerprinting workflow was carried out combining template matching strategies on the 2D-patterns of volatiles. Quantification of target volatiles (n=42) was achieved via Multiple Headspace SPME, external standard calibration and FID predicted relative response factors (RRF). HS-SPME combined to GC×GC-MS/FID and accurate quantification by predicted RFF resulted to be a great tool in the quality assessment of EVOO samples. By the effective exploration of the information encrypted in EVOOs volatilome and the accurate quantification of key-odorants an Artificial Intelligence smelling machine is realized with peculiar comparative possibilities for EVOOs aroma qualities.