Error with CladeAges calibrations with FBD model

[chaseg...@gmail.com]

Hello,

I am attempting to run a Fossilized Birth Death Model with two fossil calibrations in CladeAges. One of the calibrations is a root calibration, and the run specifies an error: 

"Cannot use originate of root. You can set useOriginate to false to fix this".

Error detected about here:

  <beast>

      <run id='mcmc' spec='MCMC'>

          <distribution id='posterior' spec='util.CompoundDistribution'>

              <distribution id='fossilCalibrations' spec='util.CompoundDistribution'>

                  <distribution id='Root.fossilprior' spec='beast.math.distributions.FossilPrior'>

If I set useOriginate="false" in the XML in the Root.fossilprior, I still receive the same error message. Wondering how to fix this problem.

Thanks for your time in advance

Chase

[Alexei Drummond]

To my mind it doesn’t make sense to mix CladeAges and the fossilized birth-death (FBD) model. The whole point of the FBD model is that you don’t need to specify any node calibrations as the FBD prior directly models the relationships between the fossil taxa and the extant taxa to provide a calibration that way. Another reason not to try to use them together is that CladeAges requires speciation, extinction and sampling rates that from a logical standpoint should be the same as the corresponding parameters in FBD but I don’t think there is any way to link them unless you assume you have perfect knowledge of those parameter values. 

I would strongly advise using one or the other and not both.

Alexei 

--
You received this message because you are subscribed to the Google Groups "beast-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beast-users...@googlegroups.com.
To post to this group, send email to beast...@googlegroups.com.
Visit this group at https://groups.google.com/group/beast-users.
For more options, visit https://groups.google.com/d/optout.

[michaelm]

Hi Chase,

I agree with Alexei's response not to mix the two approaches. But I think the error message is caused by your attempt to place a CladeAge calibration on the root node, which would not make much sense. When you place CladeAge constraints on a clade, then these always constrain the stem age of the clade, meaning the age of divergence of this clade from its sister. When the constrained clade comprises all taxa of your data set (as is the case when you constrain the root), then the tree estimated by BEAST2 does not include the sister lineage of that clade and therefore also not the stem of that clade.

In brief, when using CladeAge, one should always constrain the ages of all clades that 1) can reasonably be assumed to be monophyletic, 2) have their sister lineage represented in the phylogeny (even if it's not know which one it is), 3) are characterized by synapmorphies (or constrained geographic distributions) so that fossils could in principle be assigned to the clade directly, not only indirectly through assignment to subclades, 4) actually have a fossil record (regardless of how old that fossil record is).

Cheers,

Michael

[pitiba...@gmail.com]

Dear Alexei and Michael,

Thanks for these explanations. But, I don't understand how is it possible to use only CladeAge? In the Priors panel (Windows), we have to choose between all the speciation models : Yule, BD, FBD... (even if CA package is well installed).

So, how to do if we want to only use CladeAge?

Kind regards,

Cédric

[michaelm]

Hi Cédric,

Just use the Birth-Death speciation model, this is appropriate in combination with CladeAge constraints.

Cheers,

Michael