Smoking and/or grilling, when carried out with traditional methods involving direct contact with wood combustion fumes, is responsible for high contamination levels with carcinogenic polycyclic aromatic hydrocarbons (PAHs). The aim of this work was to investigate the PAH content of different smoked or grilled meat and fish products commonly consumed in Nigeria. A rapid method involving microwave-assisted saponification and simultaneous extraction followed by solid-phase extraction (SPE), high-performance liquid chromatography (HPLC) separation and spectrofluorometric detection was employed. Samples that were smoked or grilled using traditional systems, which use a wood fire, were heavily contaminated with benzo[a]pyrene (BaP) at levels ranging from 2.4 to 31.2 µg kg⁻¹ wet weight. Considerably lower contamination levels were found in samples smoked or grilled in the laboratory using a charcoal fire (BaP from 0.7 to 2.8 µg kg⁻¹ wet weight). The health risk associated with a daily consumption of 100 g of these products was also evaluated using the margin of exposure (MOE) approach. MOE values lower than 10,000 were obtained for all smoked/grilled commercial samples, indicating a potential concern for consumer health.

This study investigates the applicability of an iterative approach aimed at defining a chemical blueprint of virgin olive oil volatiles to be correlated to the product sensory quality. The investigation strategy proposed allows to fully exploit the informative content of a comprehensive multidimensional gas chromatography (GC×GC) coupled to a mass spectrometry (MS) data set. Olive oil samples (19), including 5 reference standards, obtained from the International Olive Oil Council, and commercial samples, were submitted to a sensory evaluation by a Panel test, before being analyzed in two laboratories using different instrumentation, column set, and software elaboration packages in view of a cross-validation of the entire methodology. A first classification of samples based on untargeted peak features information, was obtained on raw data from two different column combinations (apolar×polar and polar×apolar) by applying unsupervised multivariate analysis (i.e., principal component analysis-PCA). However, to improve effectiveness and specificity of this classification, peak features were reliably identified (261 compounds), on the basis of the MS spectrum and linear retention index matching, and subjected to successive pair-wise comparisons based on 2D patterns, which revealed peculiar distribution of chemicals correlated with samples sensory classification. The most informative compounds were thus identified and collected in a "blueprint" of specific defects (or combination of defects) successively adopted to discriminate Extra Virgin from defected oils (i.e., lampante oil) with the aid of a supervised approach, i.e., partial least squares-discriminant analysis (PLS-DA). In this last step, the principles of sensomics, which assigns higher information potential to analytes with lower odor threshold proved to be successful, and a much more powerful discrimination of samples was obtained in view of a sensory quality assessment.