Notes on the birth-death prior with fossil calibrations for Bayesian estimation of species divergence times

作者：dos Reis Mario^*

来源：Philosophical Transactions of the Royal Society B-Biological Sciences, 2016, 371(1699): 20150128.

DOI：10.1098/rstb.2015.0128

• phylogeny
• divergence time
• molecular clock
• time prior
• Bayesian method
• birth-death process

Constructing a multi-dimensional prior on the times of divergence (the node ages) of species in a phylogeny is not a trivial task, in particular, if the prior density is the result of combining different sources of information such as a speciation process with fossil calibration densities. Yang & Rannala (2006 Mol. Biol. Evol. 23, 212-226. (doi:10.1093/molbev/msj024)) laid out the general approach to combine the birth-death process with arbitrary fossil-based densities to construct a prior on divergence times. They achieved this by calculating the density of node ages without calibrations conditioned on the ages of the calibrated nodes. Here, I show that the conditional density obtained by Yang & Rannala is misspecified. The misspecified density can sometimes be quite strange-looking and can lead to unintentionally informative priors on node ages without fossil calibrations. I derive the correct density and provide a few illustrative examples. Calculation of the density involves a sum over a large set of labelled histories, and so obtaining the density in a computer program seems hard at the moment. A general algorithm that may provide a way forward is given.
This article is part of the themed issue 'Dating species divergences using rocks and clocks'.

• 出版日期2016-7-19