Abstract
The genetic basis of phenotypic or adaptive parallelism can reveal much about constraints on evolution. This study investigated the genetic basis of a canonically parallel trait: pelvic spine reduction in sticklebacks. Pelvic reduction has a highly parallel genetic basis in threespine stickleback in populations around the world, always involving a deletion of the pel1 enhancer of Pitx1. We conducted a genome-wide association study to investigate the genetic basis of pelvic spine reduction in 3 populations of brook stickleback in Alberta, Canada. Pelvic reduction did not involve Pitx1 in any of the 3 populations. Instead, pelvic reduction in 1 population involved a mutation in an exon of Tbx4, and it involved a mutation in an intron of Lmbr1 in the other two populations. Hence, the parallel phenotypic evolution of pelvic spine reduction across stickleback genera, and among brook stickleback populations, has a nonparallel genetic basis. This suggests that there may be redundancy in the genetic basis of this adaptive polymorphism, but it is not clear whether a lack of parallelism indicates a lack of constraint on the evolution of this adaptive trait. Whether different pleiotropic effects of different mutations have different fitness consequences or whether certain pelvic reduction mutations confer specific benefits in certain environments remains to be determined.