Genealogies from time-stamped sequence data

We review and develop Bayesian statistical methods for recovering genealogical structure, population size and mutation rates from radiocarbon-dated fossil mtDNA sequence data. It is possible to obtain ages for fossil DNA sequences and their common ancestors, by fitting a population-genetic model. We describe the observation model and show how uncertainty in reconstructed parameter values may be quantified via sample-based inference. We give an example, in which errors arising from radiocarbon calibration of fossil sequences are dominated by uncertainty in the genealogy and associated population parameters. We do not discuss likely model mispecification in any detail.