Can microsatellites be used to infer phylogenies? Evidence from population affinities of the Western Canary Island lizard (Gallotia galloti).

Population phylogeographic studies are generally based solely on mtDNA without corroboration, from an independent segregating unit (i.e., nuclear genes), that the mtDNA gene tree represents the organismal phylogeny. This paper attempts to evaluate the utility of microsatellites for this process by use of the Western Canary Island lacertid (Gallotia galloti) as a model. The geological times of island eruptions are known, and well-supported mtDNA phylogenies exist (corroborated as the organismal phylogeny rather than just a gene tree by nuclear random amplified polymorphic DNAs (RAPDs)). The allelic variation in 12 populations from four islands (representing five haplotype lineages) was investigated in five unlinked microsatellite loci. Analysis of molecular variance showed this data to be highly structured. A series of genetic distances among populations was computed based on both the variance in allele frequency (i.e., F(st) related) and the variance in repeat numbers (i.e., R(st) related). The genetic distances based on the former were more highly correlated with the mtDNA genetic distances than those based on the latter. All trees based on both models supported the primary division shown by mtDNA and RAPDs, which is dated at ca. 2.8 to 5.6 mybp (depending on calibration of the mtDNA clock) and which could, under the evolutionary species concept, be regarded separate species. This was achieved despite theoretical problems posed by the use of few loci, suspected bottlenecks, and large population sizes. The finer details were less consistently represented. Nevertheless, this study demonstrates that even a small number of microsatellites can be useful in corroborating the deeper divisions of a population phylogeny.