Estimating uncertainty in divergence times among three-spined stickleback clades using the multispecies coalescent

Incomplete lineage sorting (ILS) can lead to biased divergence time
estimates. To explore if and how ILS has influenced the results of a
recent study of worldwide phylogeny of three-spined sticklebacks
(Gasterosteus aculeatus), we estimated divergence times among major clades
by applying both a concatenation approach and the multispecies coalescent
(MSC) model to single-nucleotide polymorphisms. To further test the
influence of different calibration strategies, we applied different
calibrations to the root and to younger nodes in addition to the ones used
in the original study. Both the updated calibrations and the application
of the MSC model influenced divergence time estimates, sometimes
significantly. The new divergence time estimates were more ancient than in
the previous study for older nodes, whereas the estimates of younger nodes
were not strongly affected by the re-analyses. However, given the applied
MSC method employs a simple substitution model and cannot account for
changes in population size, we suggest that different analytical
approaches and calibration strategies should be used in order to explore
uncertainty in divergence time estimates. This study provides a valuable
reference timeline for the ages of worldwide three-spined stickleback
populations and emphasizes the need to embrace, rather than obscure,
uncertainties around divergence time estimates.