Estimating divergence dates and substitution rates in the Drosophila phylogeny

Obbard DJ, Maclennan J, Kim K-W, Rambaut A, O?Grady PM & Jiggins FM

(2012) Mol Biol Evol 29, 3459-3473.

An absolute timescale for evolution is essential if we are to associate evolutionary phenomena, such as adaptation or speciation, with potential causes, such as geological activity or climatic change. Timescales in most phylogenetic studies use geologically dated fossils or phylogeographic events as calibration points, but more recently it has also become possible to use experimentally-derived estimates of the mutation rate as a proxy for substitution rates. The large radiation of drosophilid taxa endemic to the Hawaiian Islands has provided multiple calibration points for the Drosophila phylogeny, thanks to the ?conveyor belt? process by which this archipelago forms and is colonised by species. However, published date estimates for key nodes in the Drosophila phylogeny vary widely, and many are based on simplistic models of colonisation and coalescence, or on estimates of island age that are not current. Here we use new sequence data from seven species of Hawaiian Drosophila to examine a range of explicit coalescent models and estimate substitution rates. We use these rates, along with a published experimentally-determined mutation rate, to date key events in drosophilid evolution. Surprisingly, our estimate for the date for the most recent common ancestor of the genus Drosophila based on mutation rate (25-40 Mya) is closer to being compatible with independent fossil-derived dates (20-50 Mya) than are most of the Hawaiian-calibration models, and also has smaller uncertainty. We find Hawaiian-calibrated dates are extremely sensitive to model choice, and give rise to point-estimates that range between 26Mya and 192Mya, depending on the details of the model. Potential problems with the Hawaiian calibration may arise from systematic variation in the molecular clock due to the long generation time of Hawaiian Drosophila compared to other Drosophila, and/or uncertainty in linking island formation dates with colonisation dates. As either source of error will bias estimates of divergence time, we suggest mutation-rate estimates be used until better models are available.

Open access, full text

Andrew Rambaut, 2007