Phylogeographic reconstructions can be biased by ancestral shared alleles: The case of the polymorphic lichen Bryoria fuscescens in Europe and North Africa.

Large phylogeographic studies on lichens are scarce, and none involves a single species within which different lineages show fixed alternative dispersal strategies. We investigated Bryoria fuscescens (including B. capillaris) in Europe and western North Africa by phenotypically characterizing 1400 specimens from 64 populations and genotyping them with 14 microsatellites. We studied population structure and genetic diversity at the local and continental scales, discussed the post-glacial phylogeography, and compared dispersal capacities of phenotypes with and without soralia. Our main hypothesis is that the estimated phylogeography, migration routes, and dispersal capacities may be strongly biased by ancestral shared alleles. Scandinavia is genetically the richest area, followed by the Iberian Peninsula, the Carpathians, and the Alps. Three gene pools were detected: two partially linked to phenotypic characteristics, and the third one genetically related to the American sister species B. pseudofuscescens. The comparison of one gene pool producing soredia and one not, suggested both as panmictic, with similar levels of isolation by distance (IBD). The migration routes were estimated to span from north to south, in disagreement with the assessed glacial refugia. The presence of ancestral shared alleles in distant populations can explain the similar IBD levels found in both gene pools while producing a false signal of panmixia, and also biasing the phylogeographic reconstruction. The incomplete lineage sorting recorded for DNA sequence loci also supports this hypothesis. Consequently, the high diversity in Scandinavia may rather come from recent immigration into northern populations than from an in situ diversification. Similar patterns of ancestral shared polymorphism may bias the phylogeographical reconstruction of other lichen species.