Dating in *BEAST

[tom]

I am using *Beast to try to confirm spp tree and branching dates. I
have 3 mitochondrial markers, thus linked, of 1155bp Cytb, 705bp COI,
and 435bp D-loop. These have 16, 5 and 34 sequences, respectively,
representing 5 spp which I have detailed in the traits section. The
differing nos. seem to present no problems - I get a spp tree with
posterior branch support of 1 or 0.95 for each branch, with expected
topology.

Less expectedly, linking the rates (each being mitochondrial) the
relative rates are Cytb>COI>D-loop in proportion 1.75:1.5:1 using
strict clock, as per the *Beast manual. This is consistent whichever
partition I set to clock rate 1.0 to estimate others. All else seems
to work well, ESS's high etc. Each partition rejects molecular clock
under ML test in MEGA 5.05, thus anticipating using relaxed
uncorrelated lognormal clock as being more realistic, or possibly
random local clock model, though these is not mentioned in the manual.
Substitution model same for each partition and as per AICc from MrAIC:
HKY+G.

In hoping to put dates to the spp tree, I have only very remote fossil
record for these freshwater fishes so can put an informative tmrca
prior on a tree node in Cytb, which I hope will carry through to the
spp tree. It does not seem to, so I could try scaling the node prior
for the other 2 gene trees by the amount that the clock rates appear
to differ in the hope that it will then appear in spp tree - any
ideas?

Scaling spp MCC tree to that node date (normal distrn. + stdev) misses
the point that spp tree dates are more recent than gene tree dates
(McCormack et al 2010 Evolution). Similarly I don't want to scale spp
tree from a guessed root age or to fix mean rate of molecular clock
model (say 2%/My?). Biogeographic evidence not much help as I'm hoping
speciation times may provide clues to dates of river captures, which
which are geologically poorly bound.

So, any ideas if my relative clock rates might be an artefact, how to
tell, & how to use cytb node tmrca to carry through to spp tree as
probably my best route to dating speciations?

I have larger datasets for D-loop of these spp which I am hoping to
use for GMRF Bayesian Skyride & more detailed biogeographic work,
incidentally, but reduced here for computation time reasons.

[tom]

Hi All,

Following on from this, I am having problems trying to apply a date
prior to an internal node in Beauti 1.6.2 (or 1.6.1) for *Beast. When
I create taxon sets under the tab, it appears to work (whether for 1
tree, or each) but then locks up Beauti such that it can't generate
the .xml file.
I can try to define my taxon sets in a text editor and then apply the
date prior, presumably for each tree.
Anyone with similar experiences, and if so, how did you work around
it?

cheers
Tom

[Hassan Ebrahimi]

Tom,

I have the same problem and I am not confident to go for editing the .xml file. Would you please explain how to edit the .xml file for defining taxon sets and setting priors to nodes? An example .xml file will be helpful too.

Your help would be appreciated.

Hassan

---

From: tom <tom.win...@gmail.com>
To: beast-users <beast...@googlegroups.com>
Sent: Tuesday, September 20, 2011 7:37 PM
Subject: Re: Dating in *BEAST

--
You received this message because you are subscribed to the Google Groups "beast-users" group.
To post to this group, send email to beast...@googlegroups.com.
To unsubscribe from this group, send email to beast-users+unsub...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/beast-users?hl=en.