Preventing fraud and assessing safety in the fishery and aquaculture sectors through rapid analytical strategies

Abstract

Mislabelling concerning species, country of origin, and method of production today represents the most frequent fraudulent practice affecting the fishery and aquaculture sectors and it is becoming a pressing problem both from an economic and a sanitary point of view. Hence, a growing need exists for tightening up fish controls on the one hand and shifting towards a risk-based control of food authenticity on the other, through the implementation of up-to-date, efficient, and standardised food inspection analytical tools, for which the requirements of rapidity, cost-effectiveness, and easiness of understanding and application are met. The present PhD dissertation was aimed at investigating the possibility of developing rapid analytical methods to address multiple authenticity issues and guarantee the integrity and safety of fish and seafood products. Fingerprinting strategies based on near infrared (NIR) spectroscopy coupled with chemometrics and profiling strategies based on inorganic elemental composition coupled with chemometrics or machine learning techniques were tested, optimised, and analytically validated to detect the mislabelling of the country of origin of different fish and seafood products. The feasibility of using one single technique to address production methods and rearing systems authenticity problems at once was also tested, together with the investigation of the possibility to concomitantly assess safety issues relating to the presence of potentially toxic metals. The following major results were obtained during the present three-years PhD program: i. Near infrared spectroscopy (NIR) and chemometric analysis can be successfully used to rapidly investigate the authenticity of European sea bass and distinguish wild from farmed fish, the intensity of the rearing system of farmed fish, and the subareas of origin in the Mediterranean Sea by performing one single spectral analysis. Specifically, the application of orthogonal partial least square discriminant analysis (OPLS-DA) to the 1100-2500 nm full scan NIR spectra of fish was found to be very useful since it allows to effectively separate non-relevant from relevant spectral information related to the specific classification purpose, thus improving the overall classification accuracy. The analysis of the NIR spectral fingerprint revealed that the variability of wavelengths associated to protein absorption was unusually significant towards sea bass discrimination based on production method/farming system. This was likely attributed to the higher influence of the muscular activity over the external feeding input of wild specimens compared to farmed ones. Instead, NIR bands associated to lipid absorption had a major contribution to identifying the geographical provenance of fish, thus suggesting further deepening of the existing relationship between fatty acid composition and environmental conditions such as water temperature and salinity. ii. NIR spectroscopy-based fingerprints of lipid and protein degradation patterns of salt-ripened anchovies was verified to be a practical mean of ensuring the traceability of fishery products after industrial processing and at different stages of the production chain. The ripening period the salted anchovy undergoes inevitably alters or modifies the original composition of fish tissue, thus masking potential markers related to the provenance. Although this, handling 1200-2500 nm NIR spectra of salt-ripened anchovies by means of OPLS-DA allowed to monitor traceability of these products, whose raw fish originates from Spanish, Tunisian, Moroccan, and Croatian fishing areas. Indeed, the discrimination by provenance of both industrial intermediate and finished anchovy products was attributed to organic fragmentation patterns generated during the manufacturing process and reflecting the original composition of the untransformed fish. An in-depth evaluation of the spectral signatures of the processed fish indicated that a complex pattern of degradation compounds deriving from proteolysis (i.e. peptides, amino acids and low-molecular-weight nitrogen compounds and other decomposition fragments) can be useful for the identification of anchovy from Morocco. On the other hand, the unsaturated lipid fraction represented a distinctive compositional mark for anchovies from Tunisia, while anchovies from Spain and Croatia were accurately identified owing to the equal contribution of protein and lipid compounds. iii. The combination of multivariate discriminant analysis with a fusion of data deriving from light stable isotope ratio of carbon and nitrogen and rare earth elements analyses was proved to be an effective strategy to detect mislabelling concerning the provenance and the method of production of sea bass. In particular, the higher isotopic abundances of carbon and nitrogen of wild-caught fish compared to aquaculture fish alone, were found to be discriminant markers for sea bass populations, thus remarking the important influence of the feeding sources and the trophic position of the fish on the final flesh composition. Since the geographical origin is known to have a more limited impact on flesh composition compared to the production method, the isotopic signature plus lanthanum and holmium concentrations (whose presence in seawater are notoriously affected by natural and anthropogenic activities) increased notably the accuracy in discrimination of fish, to a point that fish from three different fishing areas in the Mediterranean Sea were reliably discernible and mislabelling potentially identifiable. iv. The protection and promotion of Chioggia cuttlefish (an Italian traditional product for which a quality mark has been recognised), was demonstrated to be feasible through the determination of the multielement profile by inductively coupled plasma mass spectrometry (ICP-MS). The specific identification of the geographical inorganic imprint of Chioggia’s cuttlefish by multivariate discriminant data analysis and variable selection led to the conclusion that the elemental pattern linked to the geographical origin of samples is a combination of macro, trace, and ultra- trace elements which are known to be absorbed by the animals from the surrounding environment. In particular, anthropogenic elements linked to the specific production area strongly emerged as a key analytical determinant for cephalopods authenticity assessment. Of note, also concentrations of some heavy toxic metals such as cadmium and arsenic, although being so low as not to represent a potential health risk, were useful for discrimination of Chioggia’s from other common cuttlefish of different origin, thus pointing out the dual safety-authenticity advantages deriving from multielement profiling strategies. v. Classification rules to easily distinguish among high-quality transformed anchovy products from the Cantabrian Sea and other competing but lower-quality products from other countries in the Mediterranean area were built by exploiting information behind multielement patterns measured by ICP-MS. The selection of a limited number of highly informative elements measured by ICP-MS (other than those verified to be misleading included into the product and originating from the manufacturing environment) together with the concomitant creation of classification rules by applying four decision tree-based machine learning algorithms ended up being an immediate, cheap, practical, and easy-to-handy procedure to figure out how the elemental composition can help in solving many actual challenges related fish quality promotion. Specifically, the results obtained suggested that only five and six (out fifty-two) originally measured elements were sufficient to effectively authenticate Cantabrian anchovies before and after packaging, respectively. Arsenic, boron, potassium, phosphorous cadmium demonstrated to be leading markers of origin, thereby emphasizing how their measurement can represent an effective and cost-effective strategy to be applied in the primary and secondary fish production sectors. By way of conclusion, the results presented in this PhD Thesis clearly highlight the effective support in fish authenticity and safety assessment provided by analytical approaches based on fingerprinting of organic component by NIR spectroscopy on the one hand, and on profiling of the inorganic components of fish on the other. Although the proposed methodologies address problems of the same nature, it is worth noting how the field of application is supposed to be different, depending first and foremost on the fish sample characteristic and response to the technique and, secondly, the actual need of the users. For instance, rapidity, non-destructive nature, ease of use and high-throughput measurement ability, make NIR spectroscopic non-targeted approach an ideal tool for quality control operations, especially in the context of daily routine and screening analysis in the food industry. Multielement-based strategies have been experienced for much longer than spectroscopic ones, and the greater sensitivity, specificity, and precision both in qualitative and quantitative determination make them particularly suitable for use within the official controls. ICP-MS equipment for multielement analysis is more expensive and less environmentally-friendly than NIR spectroscopy instrumentations, but on the upside the lower number of samples required for proper calibration is relevant for expensive fish samples.