Home

The objective of this study was to test the possibility of using lipid profiles obtained by gas chromatography (GC) and ^sup 13^C nuclear magnetic resonance (NMR) in authentication of cod liver oils according to wild/farmed and geographical origin. GC and ^sup 13^C-NMR data of cod liver oil from wild and farmed fish from different locations in Norway and Scotland were obtained, and analyzed by principal component analysis (PCA) and linear discriminant analysis (LDA) to test if it was possible to differentiate oil from wild and cultured cod (Gadus morhua L.), and to further elucidate differences between fish from the different farms/catch area. Cod liver oils of wild and farmed origin were clearly separated in the PCA score plot both from GC and NMR data. From NMR data it was also possible to observe groupings based on geographical origin (farm/catch area) of the different samples. Using LDA with cross validation the wild/farmed classification rates were 97% for GC data and 100% for NMR data. In the classification of cod liver oils according to geographical origin (38 samples from six different farms/catch area), the correct classification rate was 63% for GC data and 95% for NMR data.

GC and NMR data: Supervised Classification LDA

For a more quantitative result on the discrimination power of the two techniques GC and ^sup 13^C NMR, LDA was performed. Both the ability to separate cod liver oil of wild from farmed fish, and to discriminate the six different groups offish (SW, NW, SF1, SF2, NF1, NF2) were investigated. Results from LDA with cross validation as validation method are given in Table 2. Wild versus farmed classifications resulted in 100% correct classification (38/38) for NMR data (two first principal components as variables). Results from GC data gave one misclassified sample and a correct classification rate of 97% (37/38) for wild/farmed classifications (three first principal components, explaining 47, 27, and 10 % of the variance respectively, used as variables). In the classification of cod liver oils according to the origin (38 samples from six different farms/catch areas), the correct classification rate was 63% for GC data and 95% for NMR data.

These results demonstrate the potential of lipid profiling by GC and ^sup 13^C NMR as authentication methods of cod liver oils according to wild/farmed origin. When it comes to the geographical origin (farm/catch area), ^sup 13^C NMR seems better able to separate oils of different groups in this study. Compared to GC, ^sup 13^C NMR is a less sensitive technique, however ^sup 13^C NMR can be applied nondestructively and without extensive sample manipulation. In addition, ^sup 13^C NMR provides a more complete picture of the lipid profile than GC, since the positional distribution of fatty acids in triacylglycerols and information about lipid classes can be obtained. The technique is suited to detect if chemical modifications have been made to the natural oil (i.e addition of diacylglycerols or ethyl-esters) [14, 15] and to study deterioration and oxidation [29]. The future goal is to establish robust databases with authentic material, so that unknown samples can be statistically treated and grouped according to origin (species, production method, geographical origin). The result, a reliable authentication method of marine oils, would be important to secure correct labeling, to discourage commercial fraud and to prevent illegal capture.