Direct Analysis in Real Time-High Resolution Mass Spectrometry as a potential tool to discriminate Farmed versus Wild type salmon

Introduction Food authenticity aims to guarantee product safety, transparency and protection of consumers' health. In order to safeguard food authenticity, a variety of analytical techniques have been investigated some of which are routinely used. If on one hand some methodologies typically based on MS detection revealed successful for the identification of selected molecules as suitable markers for species identification, on the other hand the untargeted methods utilizing MS detection appear not deeply investigated in this specific field. In the present communication, a non targeted approach based on DART ionization coupled to High Resolution Mass Spectrometry was described and applied for the discrimination of two types of salmons -farmed versus wild type-, belonging to the Salmo salar species and collected in Canada. In particular, the apolarfractions were analysed and the final data were further subjected to multivariate statistical analysis to highlight eventual differences between bothdata-set. Experimental Different fish homogenates were prepared from 26 canadian wild type salmons (CND-WT) and25 canadian farmed salmons (CND-F). Aliquots of 2.5 g of each fish homogenate were added with 5 mL of refrigerated methanol and 5 mL of refrigerated chloroform and stirred for 15 min at R.T.After decantation for 2 hours at 4°Cthe lower apolar fraction was withdrawn and first evaporated under nitrogen flow and then resuspended in an equal volume of isopropanol before HRMS analysis. Analysis were performed by DART ionization (IonSense) coupled to HRMS using an Exactive(TM) (Thermo Fisher Scientific) system in negative ion mode by setting the gas temperature 150°C. The total peak lists of all ions generated from each analysis was converted in .csv format and processed for the multivariate statistical analysis by using the free online software MetaboAnalyst 3.0. Results After duly inspecting the HRMS spectra referred to the analysis of the apolar fraction, it appears that the ion m/z 281.247 is the most intense in CND-F samples despite the ion m/z 255.232 instead dominating the CND-WT spectra. By entering the m/z values into the LipidMaps database, the ion species were likely attributed to thestearic acid (m/z 281.247) and palmitic acid (m/z 255.232) suggesting their possible role as potential biomarkers indicator of the type of capture. By applying the multivariate statistical analysis e.g. the Principal Component Analysis(PCA) on the list of ionsa good separation between the two fish groups (farmed vs wild type) was foundwith PC1 and PC2 that explained respectively 68.5% and 23.3% of the total variance. Conclusions Ion differences between the CND-F and CND-WT samples werehighlightedby applying DART-HRMS analysis followed by multivariate statistical analysis on the whole dataset. In particular, PCA resultsshowed a good separation of both salmon groups. In future, a supervisioned approach might be envisaged to further improve such separation between both groups.