Does organismal pedigree impact the magnitude of topological congruence among gene trees for unlinked loci?
- 74 Downloads
One of the fundamental assumptions in the multi-locus approach to phylogeographic studies is that unlinked loci have independent genealogies. For this reason, congruence among gene trees from unlinked loci is normally interpreted as support for the existence of external forces that may have concordantly shaped the topology of multiple gene trees. However, it is also important to address and quantify the possibility that gene trees within a given species are all inherently constrained to some degree by their shared organismal pedigree, and thus in this strict sense are not entirely independent. Here we demonstrate by computer simulations that gene trees from a shared pedigree tend to display higher topological concordance than do gene trees from independent pedigrees with the same demographic parameters, but we also show that these constraining effects are normally minor in comparison to the much higher degree of topological concordance that can routinely emerge from external phylogeographic shaping forces. The topology-constraining effect of a shared pedigree decreases as effective population size increases, and becomes almost negligible in a random mating population of more than 1,000 individuals. Moreover, statistical detection of the pedigree effect requires a relatively large number of unlinked loci that far exceed what is typically used in current phylogeographic studies. Thus, with the possible exception of extremely small populations, multiple unlinked genes within a pedigree can indeed be assumed, for most practical purposes, to have independent genealogical histories.
KeywordsDispersal Gene flow Genealogy Phylogeny Phylogeography Vicariance
We thank J. Kissinger, the Institute of Bioinformatics, and the Research Computing Center at University of Georgia for providing computation resources. Comments from anonymous reviewers and editor greatly improved this manuscript. C.-H.K. was supported by a National Institute of Health Training Grant (GM07103) and the Alton Graduate Research Fellowship to the Department of Genetics at University of Georgia, and J.C.A. was supported by funds from the University of California at Irvine.