772
Experimental Designs for Genetic Parameters and Response to Selection of Social Interaction Traits in Aquaculture
Variance component estimation of social interaction traits requires a large number of small groups (250 to 500) of two families, which may not be practical in aquaculture. Stochastic simulations were performed to study alternative experimental designs. Three alternative designs were compared with respect to variance component estimates and response to selection, with groups composed: at random, of two families and of three families. In each design, the number of groups approximated 100. Groups composed of two and three families with a group size of at least 30 individuals were statistically more powerful to estimate direct and social variance components than groups composed at random. Response to selection from groups of two and three families was substantially higher than the response from groups composed at random.
Keywords:
Experimental design
Genetic parameter
Response to selection
Socially affected trait
Stochastic Simulation
