How to genetically increase fillet yield in fish: Relevant genetic parameters and methods to predict genetic gain

Highlights

• Genetic parameters of fillet yield are not predictive of genetic gain.

• Selection index theory enables efficient prediction of genetic gain in fillet yield.

• Genetic parameters of residual fillet weight are reasonably predictive.

• We provide recommendations for efficient prediction of genetic gains in fillet yield.

• Current published literature mostly does not comply with these recommendations.

Abstract

Fillet yield (i.e. the proportion of edible muscle in a fish) is a key economic trait for species sold as fillets. Its genetic improvement is complicated by several of its characteristics 1) it is a ratio trait, 2) its numerator (fillet weight) and denominator (body weight) are strongly correlated (correlations in the range 0.89–0.99), 3) it offers little phenotypic variation and 4) it cannot be measured on alive breeding candidates. In a former study, we showed that it could be improved by selection, especially with three selection indices, fillet yield, residual fillet weight and a ratio-specific linear index. However, it is well known that the heritability of ratio traits does not permit a reliable prediction of genetic gains. As predictability of genetic gains is a key requirement to define breeding programs, we investigated how genetic gains in fillet yield could be predicted by the genetic parameters of fillet yield, of residual fillet weight and of the component traits of the linear index. To this end, we compared simulated genetic gains with those estimated by classical prediction methods. This was done using real sets of genetic parameters obtained in nine populations of rainbow trout, European sea bass, gilthead sea bream and common carp. We show that the genetic parameters of fillet yield cannot be used to reliably predict genetic gains in fillet yield. Conversely, selection index theory using a linear index, combining either fillet weight and body weight or fillet weight and waste weight, provides almost perfect prediction of gains. Still, it is highly sensitive to the precision of the genetic and phenotypic correlations estimates, which should not be rounded to less than three decimals for fillet weight and body weight, while two decimals are appropriate for fillet weight and waste weight. A simple, reasonably precise alternative to the linear index is the use of residual fillet weight (the residual of the regression of fillet weight on body weight) as a surrogate for fillet yield.

Introduction

In fish selective breeding programs, the initial focus for selective breeding has always been growth rate (Chevassus et al., 2004; Gjedrem, 2012). However, the value of the round weight gain obtained is not the same in fish with high or low fillet yield (fillet weight/body weight ratio). This value can easily be turned into economic gains for species sold as fillets, where fillet yield can have a large impact on value added and net profit (Kankainen et al., 2016). Increasing the edible part of fish is also expected to decrease the environmental impact of the production of a given amount of edible fish flesh (Acosta Alba et al., 2015). The same reasoning applies for similar traits in other aquatic species, such as tail percentage in shrimp (Campos-Montes et al., 2017) or meat yield in shellfish (Nguyen et al., 2011).

Selective breeding on a ratio is seen as a problematic issue, which has been studied in many farmed animals. The main focus has been given to feed conversion ratio (FCR), the ratio of average daily feed intake to average daily gain, which has a major economic impact in all intensive farming systems. Most studies about selection methods for ratio traits have thus focused on FCR (Famula, 1990; Gunsett, 1984; Lin, 1980; Lin and Aggrey, 2013; Varkoohi et al., 2010). The general conclusion of these studies is that selection on a linear index combining the numerator and the denominator trait is generally more efficient than direct selection on the ratio or on one of its component traits. We recently demonstrated by simulation that although the numerator and denominator of fillet yield are very highly correlated (genetic correlations in the range 0.93–0.99), selection for fillet yield should be efficient, albeit with moderate gains in the range 0.30–0.95% fillet per generation (Fraslin et al., 2018). This may be achieved with different selection indices, among which fillet yield itself, residual fillet weight (the residual of the regression of fillet weight on body weight) or a linear index combining fillet weight and waste weight, the latter being defined as the difference between body weight and fillet weight. In this previous study, we did not test a linear index combining fillet weight and body weight, but although these traits are more strongly correlated than fillet weight and waste weight, it would be plausible that linear index selection also works with such an index.

The fact that selection gain on fillet yield can be obtained by those methods does not imply however, that such selection gains can easily be predicted from their genetic parameters. Indeed, it was previously shown that the heritability of a ratio trait, estimated from the analysis of covariance between relatives, was substantially different from the estimate obtained from simulated genetic gains (Gunsett, 1987), thus showing that the heritability of the ratio cannot be used to predict genetic gain. The same study showed that a method using selection index theory to approximate the process of selection on a ratio was more efficient, but still did not provide an exact estimate of genetic gain in all situations. On the contrary, when selection is done not on the ratio but on a linear index combining the numerator and the denominator in an optimal way, as proposed by Lin (1980) the genetic gain is expected to be perfectly predictable using standard selection index theory as long as the heritability, genetic and phenotypic correlations, and the phenotypic variance of both component traits are known (Lin and Aggrey, 2013).

In general, prediction of gain on a single trait is very simple, as it can be done (in the case of mass selection) using the classical breeder's equation ∆_G = ih²σ_P, which just requires the knowledge of the heritability h², the selection intensity i and the phenotypic standard deviation of the trait σ_P to estimate the genetic gain Δ_G (Falconer and Mackay, 1996). On the contrary, predicting gain from a two-trait linear index requires h² and σ_P for the two component traits, as well as the phenotypic and genetic correlation between them. In addition, it requires a bit of matrix algebra, versus a simple multiplication of three terms.

If we want to predict genetic gain when using a linear index aimed at increasing the fillet to waste ratio, as suggested in Fraslin et al. (2018), the genetic parameters of fillet weight and waste weight are not available in the fish breeding literature today. If we want to predict genetic gain using a linear index aimed at increasing fillet yield, the genetic and phenotypic correlations of fillet weight and body weight are probably not known, or at least published, with sufficient precision (e.g. 0.99 can be anything from 0.9850 to 0.9949, which makes a big difference for such values which are very close to unity). In the published literature, most studies give genetic parameters of fillet yield, and quite a number also genetic parameters of residual fillet yield or log-residual fillet yield (e.g. Haffray et al., 2012; Prchal et al., 2018; Vandeputte et al., 2017; Vandeputte et al., 2014). Thus, assessing which parameters to use in order to predict genetic gain in fillet yield is important 1) to assess which information from the existing literature is usable to predict genetic gains efficiently and 2) to eventually propose new sets of parameters to estimate in genetic studies of fillet yield in fish.

To this end, we estimated the genetic parameters of body weight, fillet weight and waste weight (defined as waste weight = body weight – fillet weight) in nine field datasets from four important aquaculture species (rainbow trout Oncorhynchus mykiss, European sea bass Dicentrarchus labrax, gilthead sea bream Sparus aurata and common carp Cyprinus carpio). In the same datasets, we also estimated the genetic parameters of fillet yield and of residual fillet weight. Then, we performed stochastic simulation of selection using different selection indices to determine the expected genetic gain in the first generation, and compared this simulated gain with estimates derived from the genetic parameters of the traits using standard quantitative genetics theory for single traits (fillet yield and residual fillet weight) and linear indices.