Allele elimination recalculated: nested subset analyses for molecular biogeographical data
Correspondence: Jan Christian Habel, Department of Ecology and Ecosystem Management Technische Universität München Hans-Carl-von-Carlowitz-Platz 2 D-85350 Freising-Weihenstephan, Germany
Post-glacial colonization of species from low-latitude refugia to high latitudes, or from lower to higher elevations, often involves repeated founder effects due to stepwise colonization. This may cause repeated population bottlenecks and the subsequent loss of alleles. Regression analyses have traditionally been used to analyse the correlation between the mean numbers of alleles and geographical distances from refugia. Here, we describe and evaluate the performance of nested subset analyses for detecting allele elimination.
Genetic data sets from five butterfly and one beetle species were reanalysed using regression and nested subset analyses.
The data sets analysed here showed both congruent and divergent results under regression and nested subset analyses. Some data sets did not feature a significant correlation between the mean number of alleles and the colonization trajectory, but did show significant nested structure. Others showed the opposite effects. Using allele frequencies from the same data sets, we did not obtain significant patterns of nestedness.
Our results indicate that classical regression analyses are not always a suitable tool for analysing allele elimination, and nestedness analyses are much more meaningful. Local natural selection can alter allele frequencies, thereby erasing biogeographical patterns that have evolved as a result of the stochastic processes involved in colonization. Thus, an appropriate means of documenting allele elimination sensu Reinig is the joint application of nested subset and regression analyses based on presence/absence and abundance data for genetic diversity.