Up-scaling validation of a dummy regression approach for predictive modelling the fillet fatty acid composition of cultured European sea bass (Dicentrarchus labrax)

Abstract

The aim of the study was to validate a dummy regression approach for predictive modelling the fillet fatty acid (FA) composition of cultured European sea bass with dietary FA composition and lipid fillet content as independent variables. The model used our own data on gilthead sea bream as reference subgroup dataset and data from turbot, sole and European sea bass as dummy variables. Most of the observed variance within and among species was explained by the regression model without statistical significant interactions on blocks between diet composition and fish species subgroups. For the validation of European sea bass FA descriptors, predictive values derived from data on fish reared at laboratory scale were plotted against those obtained in farmed fish harvested at commercial size. A close linear association near to equality was found for 12 representative FAs, including saturated FAs, monoenenes and polyunsaturated FAs. This finding reinforces the possibility to produce tailored and healthy seafood products according to the guidelines of essential FA requirements in humans. FA algorithms for all the species in the model are hosted at www.nutrigroup-iats.org/aquafat as a multispecies tool to interrogate the nutritionally regulated FA composition of four cultured marine fish species of a high added value.