Tracing evolution through space and time
Beth Shapiro, The Pennsylvania State University
Since its inception, the field of “Ancient DNA” has promised to make it possible to watch the process of evolution within populations as it actually occurs. However, limitations in the statistical analysis of serially-sampled sequences (data isolated from individuals of different ages) have made this difficult to achieve. Recent advances have made it possible to reconstruct changes in genetic diversity within populations, and to correlate these with particular environmental events, such as the peak of the last ice age, and to measure directly the effect of large-scale environmental change on populations. A major limitation, however, has been to incorporate spatial as well as temporal information in these analyses. Inference of spatial patterns in the data, such as may result from long-range dispersal or local population extinctions and replacements, has been limited for the most part to post-hoc analysis of one or a set of phylogenetic trees. I will discuss recent advances in statistical phylogenetics that allow spatial as well as temporal demographic patterns to be directly inferred from serially-sampled sequence data by co-estimating evolutionary and demographic parameters within a likelihood framework. Using data from several large mammal species that were once abundant in Beringia, an ice-free Pleistocene refuge, I will illustrate how these methods can be used to reconstruct more than 150,000 years of evolution within populations.