Limited genetic parallels underlie convergent evolution of quantitative pattern variation in mimetic butterflies

Mimetic systems allow us to address the question of whether the same genes control similar phenotypes in different species. Although widespread parallels have been found for major effect loci, much less is known about genes that control quantitative trait variation. In this study, we identify and compare the loci that control subtle changes in the size and shape of forewing pattern elements in twoHeliconiusbutterfly co-mimics. We use quantitative trait locus (QTL) analysis with a multivariate phenotyping approach to map the variation in red pattern elements across the whole forewing surface ofHeliconius eratoandHeliconius melpomene. These results are compared with a QTL analysis of univariate trait changes, and show that our resolution for identifying small effect loci is somewhat improved with the multivariate approach, but also that different loci are detected with these different approaches. QTL likely corresponding to the known patterning geneoptixwere found in both species but otherwise, a remarkably low level of genetic parallelism was found. This lack of similarity indicates that the genetic basis of convergent traits may not be as predictable as assumed from studies that focus solely on Mendelian traits.