Comparison of designs for estimating genetic parameters and obtaining response to selection for social interaction traits in aquaculture

Abstract

Social interactions among individuals may affect individual productivity and welfare in aquaculture. Since social effects may have a genetic component, known as an indirect genetic effect (IGE), genetic selection may be a promising tool to simultaneously improve welfare and productivity in aquaculture. Here we compare two experimental designs that have been previously proposed for the genetic improvement of socially affected traits. We used stochastic simulations to compare a design where each group consisted of members of two families (2-FAM) with a design where each group consisted of members of three families (3-FAM). The 2-FAM and 3-FAM designs were compared using an equal number of groups (96 groups). The group size, i.e., number of individuals within each group, was either 30 or 60 individuals. Both designs were compared for the precision of estimated direct and social genetic parameters and for response to selection, either with or without a restriction on the rate of inbreeding. Four different schemes with a low variance (heritability for social effects or hS2=0.1), a moderate variance (hS2=0.3), or a high variance (hS2=0.5) of the social genetic effects, and a negative correlation between direct and social genetic effects (rAD,S=−0.6) were compared. The negative rAD,S indicates competition between group mates. Differences in precision of estimated genetic parameters between both designs were small. At low hS2 and group size of 30, the 2-FAM design was superior with respect to precision of social additive genetic variance. When the social genetic variance was small, a larger group size, e.g. 60 is recommended. The 2-FAM design resulted in a higher accuracy of selection for social and total genetic effects, but also in a higher rate of inbreeding compared to the 3-FAM design. When the rate of inbreeding was restricted to ~2%, the total response to selection was significantly higher for the 3-FAM design. In conclusion, the 2-FAM and 3-FAM designs differ little in accuracy of the direct and social genetic parameter estimates, while the 3-FAM design is superior with respect to the response to selection at a fixed rate of inbreeding. This study is particularly important for making decisions regarding experimental design when breeding for social genetic effects in aquaculture. Statement of relevanceCommunal or family rearing with a large number of animals in a group is common practice in aquaculture. Large variation in body size, which may inflate competition for feed, is generally observed in the communal rearing tank. In addition, cannibalistic or aggressive behaviour is frequently observed in a large variety of fish species. So far, the issue is addressed by management measures, such as size grading, which may be labour intensive. Sustainable solutions through selective breeding have not been tested.Previous studies compared two experimental designs for social interaction traits; groups composed of two families (2-FAM) and three families (3-FAM) with groups composed at random and found that the 2-FAM and 3-FAM designs were better with respect to genetic parameter estimates than designs with groups composed at random. It is however still unknown whether the 2-FAM or the 3-FAM design is better. Previous studies did not consider the rate of inbreeding when studying the designs. This paper describes the results from a comparison of the two experimental designs for estimating genetic parameters and obtaining response to selection, taking into account the rate of inbreeding. As such, this paper should be of interest to a broad readership including those interested in aquaculture genetics.