Prediction of fillet fatty acid composition of market-size gilthead sea bream ( Sparus aurata) using a regression modelling approach

Abstract

Gilthead sea bream ( Sparus aurata) were fed in triplicate groups with a commercial standard diet from the juvenile stage to male–female sex reversal under natural day-length and temperature conditions. Every 3–4 months during the two-year production cycle, 9 fish were randomly selected and sampled for flesh composition analyses of total lipid levels and fatty acid (FA) composition. Curvilinear regressions fitting total lipid levels and % FAs in total lipids were made to underline the differential distribution of a given fillet FA within neutral and polar lipid fractions. This dataset along with published results on market-size fish were combined for multilinear regression approaches, with the aim of describing strong relationships (P < 0.0001) between fillet FA composition and two independent variables: dietary FA composition and fillet lipid level. For saturated (14:0, 16:0, 18:0) and monounsaturated (16:1n-7, 18:1n-7, 18:1n-9, 20:1n-9) FAs, the overall variance in fillet FA composition is primarily explained by dietary FA composition and secondly by fillet lipid level. This second independent variable also contributes to explain the variations observed in arachidonic acid (20:4n-6) and docosahexaenoic acid (22:6n-3), but a statistically significant contribution is not found for linoleic acid (18:2n-6), linolenic acid (18:3n-3), eicosapentaenoic acid (20:5n-3) and docosapentaenoic acid (22:5n-3). The consistency of these predictive equations in our particular rearing conditions was proved by means of a test validation trial, using fish fed an experimental diet based on plant proteins and fish oil.