Abstract
AbstractAccurately measuring individual feed intake is required to include feed efficiency (FE) as an objective in commercial breeding programs. Phenotyping individual feed intake through direct measurements remains complex in fish reared in groups. One way to overcome this challenge is to find proxies for estimating FE. This study aimed to investigate the correlations between fish FE and potential predictive criteria in rainbow troutOncorhynchus mykiss. As predictive criteria, we considered the variations of body weight assessed as thermal growth coefficients (TGC) and the feeding behaviour assessed as the number of feed demands over a period and the proportion of the demands made before noon. Feed intake was assessed over three different periods in ten isogenic lines allowing a recording for each of the ten genotypes while keeping fish in groups. The protocol consisted of two successive feed deprivation and refeeding phases after initial basal growth periods. Correlations were then calculated between FE, measured either as feed conversion ratio (FCR) or residual feed intake (RFI), and the different indirect criteria. We found positive phenotypic correlations between FCR and RFI over the feed intake measurement periods (r ⩾ 0.79,P< 0.001). Assessment of the relationship between FE traits (FCR and RFI) calculated over the three measurement periods and TGC revealed no significant association. We found significant positive correlations between RFI calculated from the first feed intake measurement period and feeding behaviour, assessed as the number of feed demands (r = 0.42-0.49,P⩽ 0.022). Nevertheless, such correlations were not established for the two other measurement periods following feed deprivation. While we demonstrated that the weight variations during feed deprivation are not a good proxy for FE (FCR or RFI) in rainbow trout, we also highlighted the interest in exploring more the relationships between FE and feeding behaviour in fish.
Publisher
Cold Spring Harbor Laboratory