Meeting Abstract

22.6  Sunday, Jan. 4  Evolution of modularity: selection for trait disassociation PAVLICEV, M*; CHEVERUD, JM; WAGNER, GP; Washington University, St. Louis; Washington University, St. Louis; Yale University, New Haven pavlicev@pcg.wustl.edu

Genetic variance in intertrait relationship provides potential for evolutionary change in integration of phenotypic units and can supply a model for the evolution of modularity. In modular genetic architecture the pleiotropic effects are restricted to traits with common development and/or function. As the mutational effects of underlying genes upon unrelated traits thus are limited, the probability of deleterious mutation is reduced, increasing evolvability. However, the origin, as well as any rewiring of modular genetic architecture by selection requires variation in the range and strength of pleiotropic effects. Recent studies have demonstrated the existence of such genetic variation by mapping so-called relationship QTL (rQTL), manifesting genetic variation in trait covariance. It was demonstrated that these loci are involved in epistatic interactions that affect the traits differently (differential epistasis; Cheverud et al. 2004; Pavlicev et al. 2008). The variation in intertrait relationship can be present even if the genotypic mean values are not affected, i.e., if the rQTL does not affect the trait means. In such cases the effect can stem from differential canalization of variance in the two traits, or from their genotype-dependent covariance. So far it was shown that stabilizing selection increases integration between traits, regardless of whether the selection itself is correlative (Jones et al. 2007). However, we lack a model for trait disintegration to enable individuation of organismal parts. Here we consider the selection on rQTL and derive the first population genetic model that allows for the selection for trait disassociation, a mechanism that may lead to disintegration of modules.