[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]

Re: Genes show Neoaves branching before K/Pg extinction

Here's the paragraph about calibration:

"Because the 10-gene dataset had broader taxon sampling and very small amounts 
of missing sequence we used it to estimate divergence times among taxa. Node 
ages were estimated with the lognormal relaxed clock method in BEAST v. 1.6.2 
[34], and were compared with relaxed clock estimates with correlated and 
uncorrelated rates in MCMCtree [35]. To calibrate molecular clocks we used 
seven fossil anchor points: crocodile/bird split 243–251 Ma [36], 
alligator/caiman split 66–71 Ma [36], emu/cassowary split 25–35 Ma [37], magpie 
goose/greater white-fronted goose split 66–68 Ma [38], penguin/petrel split 
60–62 Ma [39], hummingbird/swift split 47–49 Ma with an upper limit of 53 Ma 
[12] and jacana/painted snipe split 30–32 Ma [40]. A uniform prior was placed 
on the origin of modern birds from 67 to 133 Ma for the relaxed clock methods. 
The lower limit was set by the oldest diagnostic neornithine fossil *Vegavis* 
[39]. The upper limit was set by mitochondrial estimates of!
 modern bird origins [9]. Multiple independent runs of 100 million generations 
were performed in BEAST with sampling every 2000 generations. Dating in 
MCMCtree was performed with the available HKY model for 100 000 generations and 
sampling every 10 generations. Dates were averaged across the runs. Convergence 
of runs in both methods were checked using Tracer (ESS > 200)."

So... only two calibration dates had older limits as well as younger ones: the 
hummingbird/swift split and the origin of Neornithes. The latter has a uniform 
prior, so every date within the huge span of time "from 67 to 133 Ma" ago was 
considered equally likely, and the older limit is _itself_ the result of 
molecular divergence dating, it is _not_ taken from the fossil record. This is 
expected to lead to overestimated divergence date estimates.

The distribution of the calibration points is much better than it could be, but 
it's still not good: the ones inside Neornithes have a rather small span of 
ages; there are only two outside it, one being about as old as the oldest one 
inside Neornithes and the other being four to five times as old. This, too, 
probably leads to overestimated divergence date estimates.