Simulation for the Investigation of a Social Model for Indirect Genetic Effects in Aquaculture
Abstract
Aquaculture is growing quickly as a method for farming aquatic based organisms for use in human consumption, and it is likely that it will produce over 60\% of total fish used in human consumption by 2030. This growth will require productivity increases. Unlike land animals fish raised in a farming environment are in very tight quarters resulting in competitive interactions. An important question is how to model this indirect genetic effect and how well it can be estimated compared to well understood direct genetic effects. An experiment with fish was performed at Dalhousie University to investigate a model including direct and indirect genetic effects. This work investigates the model proposed by Peter Bijma, explaining its background statistically and biologically and then performing a simulation study to determine how well this model behaves under various circumstances, namely different possible values of parameter sets for the model, and to see how well it estimates the variances involved in a situation deemed optimal by the model proposer. Interesting results are seen in that the modelling of the covariance is more complicated than expected, and that the direct genetic variance is more easily accurately estimated than the indirect genetic variance, as would be expected in a biological context.