Resolving Species Phylogenies of Recent Evolutionary Radiations1

Abstract

For recently derived species and when the time separating speciation events is short, the phylogenetic distribution of taxa in a gene tree may not accurately reflect the actual species relationships. The phylogenetic tradition of relying on gene tree and species tree synonymy is not reliable under such historical scenarios. Nevertheless, recent studies have demonstrated that accurate estimates of species relationships are possible when the method of phylogenetic inference considers not only the stochastic processes of nucleotide substitution, but also the random loss of gene lineages by genetic drift—even when there is widespread incomplete lineage sorting. This simulation study examines how the broader phylogenetic context, that is, the species tree topology and branch lengths, influences the ability to recover species relationships when taxa have undergone a recent and rapid radiation. As expected, the time since species divergence and the time between speciation events influences whether phylogenetic relationships are accurately estimated. However, the influence of the timing of divergence on the ability to recover species relationships accurately differed depending on the relative position of the taxa in the species tree. Differences in the ability to recover these relationships across multiple simulated species trees highlight the potential effects of taxon sampling on phylogenetic inference at, or near, the species boundary and under high rates of speciation. By focusing attention on the species tree, rather than on the individual gene trees as the basis for interpretations about species relationships, these results also represent a fundamental shift from the phylogenetic paradigm.