BEAST terminated with an error

[Yugh]

Dear Beast users:

I created a XML file in Beauti to estimate divergence date. It is
strange that sometimes this Beauti file can be executed by Beast, but
sometimes it was terminated with an error. The following is the error
information warned by Beast:

The initial model is invalid because state has a zero probability.
If the log likelihood of the tree is -Inf, his may be because the
initial, random tree is so large that it has an extremely bad
likelihood which is being rounded to zero.
Alternatively, it may be that the product of starting mutation rate
and tree height is extremely small or extremely large.
Finally, it may be that the initial state is incompatible with one or
more 'hard' constraints (on monophyly or bounds on parameter values.
This will result in Priors with zero probability.
The individual components of the posterior are as follows:
The initial posterior is zero:
CompoundLikelihood(compoundModel)=(
BooleanLikelihood=0.0,
DistributionLikelihood=-2.7081,
DistributionLikelihood=-2016.579,
DistributionLikelihood=-Inf,
DistributionLikelihood=-25.8718,
DistributionLikelihood=-Inf,
DistributionLikelihood=-1.8654,
DistributionLikelihood=0.0986,
OneOnX(constant.popSize)=-4.3175,
CoalescentLikelihood(coalescentLikelihood)=-1029.6749),
CompoundLikelihood(compoundModel)=(
TreeLikelihood(treeLikelihood)=-27204.8071).

How to fix this problem? Many thanks.

Yu

[Hassan Ebrahimi]

Dear  Yu,

This is a common problem and it happens because BEAST starts with a randomly generated tree and if you set tight priors for the parameters, the starting tree may have an extremely low probability which makes it impossible for BEAST to proceed. To my knowledge, one solution to this problem is to define a proper "user_specified" starting tree instead of a randomly generated one. This option is available in BEAUti panel under Tree tab. Again you may have no idea of a proper starting tree for your data. In this case you can once run BEAST with wider priors for some tens of thousands of MCMC chain and then take the Maximum Clad credibility tree identified by TreeAnnotator as the starting tree for the next run which is for your actual prior settings. This trick worked for me, I hope it works for you too.

Regards,

Hassan

---

From: Yugh <yugh...@gmail.com>
To: beast-users <beast...@googlegroups.com>
Sent: Friday, December 23, 2011 4:09 PM
Subject: BEAST terminated with an error

--
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.

[Diogo Silva]

Dear Yugh,

Besides what was already mentioned by Hassan, did you enforced any
monophyly constraints? If so, the user specified tree must be in
accordance with that constraint, otherwise the 0 likelihood error will
appear. If more than 2 contraints were enforced it also advisable that
you double-check them in order to avoid any potential conflicts that
will also produce the same error.

Cheers
Diogo

[Yugh]

Thanks for the replies of Hassan and Diogo. I enforced several
monophyly constraints calibrated by fossile records or previous
molecular datings. If the tree identified by TreeAnnotator includes
a portion of these constraints and I take it as the starting tree for
the next run, the other constraints can not be enforced in the next
run? In addition, is it possible that calibrations are incompatible?
Thanks.

Regards,
Yu