Abstracts (first author)
Assessing the confounding effect of population structure on Bayesian skyline plot inferences of demographic history
It is well known that population structure can confound coalescent-based inferences of past population size changes when it is not properly accounted for. Although the vast majority of species are genetically structured to some extent, few studies have quantified how different types and levels of structure might confound demographic inference. Among the most widely applied demographic inference methods is the Bayesian skyline plot (BSP and the derived EBSP), which assume a panmictic model without structure. We simulated DNA sequence data under a variety of scenarios involving structured populations with variable levels of gene flow and analysed them using EBSPs as implemented in the software package BEAST. Results revealed that BSPs can show false signals of population decline under several biologically plausible combinations of population structure and sampling strategy, suggesting that the interpretation of several previous studies may need to be re-evaluated. We found that sampling strategies present a trade-off between minimizing the risk of false positives and maximizing the power to detect recent demographic events. Therefore, a balanced sampling strategy whereby samples are distributed over several populations provides the best scheme for inferring demographic change over the time scale that is typically of interest. Case studies demonstrate that although the structure effect should always be accounted for, some systems will be more susceptible than others. Our study provides a much-needed quantification of the structure effect in BSP analyses and some practical guidelines to keep in mind before attempting such analyses.