The influence of taxon sampling and tree shape on molecular dating: an empirical example from Mammalian mitochondrial genomes.

Abstract

Over the last decade, molecular dating methods have been among the most studied subjects in statistical phylogenetics. Although the evolutionary modelling of substitution rates and the handling of calibration information are the primary focus of species divergence time research, parameters that influence topological estimation, such as taxon sampling and tree shape, also have the potential to influence evolutionary age estimates. However, the impact of topological parameters on chronological estimates is rarely considered. In this study, we use mitochondrial genomes to evaluate the influence of tree shape and taxon sampling on the divergence times of selected nodes of the mammalian tree. Our results show that taxon sampling affects divergence time estimates; the credibility intervals for age estimates decrease as taxonomic sampling increases (i.e., estimates become more precise). The influence of taxonomic sampling was not observed on nodes that lay deep in the mammalian phylogeny, although the means of the posterior distributions tend to converge with increased taxon sampling, an effect that is independent of the location of the node. In the majority of cases, the effect of tree shape was negligible.