Meeting Abstract
96.7 Monday, Jan. 6 15:00 Genetic modeling to study limits in selection experiments for behavioral traits HIRAMATSU, L.*; WOLAK, M.; GARLAND, T.; Univ. of California, Riverside; Univ. of California, Riverside; Univ. of California, Riverside lhira001@ucr.edu
Many behaviors can be viewed as composite traits because the total amount of expression is the product of intensity and duration. Each of these two components is, in turn, potentially limited by either the organism’s physical ability or its level of motivation. Therefore, even if all alleles at every locus affecting physical ability and/or motivation have purely additive effects, the way they interact to determine intensity or duration of a behavior is non-additive. If selection is performed on a composite behavioral trait with such properties, then non-additive interactions (dominance and/or epistasis) could be important in determining the response to selection. We created a simple model with purely additive effects and the above interactions to simulate long-term selection experiments. Population sizes were chosen to mimic typical selection experiments with rodents (Ne=40/generation) and 40% of individuals were selected per generation. Overlap in the phenotypic effects of the loci on the four lowest-level traits was varied, as were starting allele frequencies. Values for allelic effects, environmental effects, and population means in the base population and when a limit might be reached were chosen to mimic a long-term selection experiment that targeted voluntary wheel running in mice (Swallow et al. 1998; Careau et al. 2013). Similar to results from the wheel-running selection experiment, the model reached a selection limit after ~20 generations, but the limit coincided with depleted additive genetic variance, unlike the selection experiment. These results suggest that the maintenance of additive genetic variance in the selection experiment cannot be explained by purely additive effects, even with complex interactions among behavioral components. Supported by NSF IOS-11212732.
