discrete phylogeography in beast2 - error initialize

[sara]

Hi,

trying to run discrete phylogeographic analyses in beast (2.4.2), adapting from tutorial "ancestral Reconstruction/Discrete Phylogeography with BEAST2.3.x"

tried to set the trait clock model both as identical to the tutorial (strict) and also similar to the other partition (relaxed logNormal) - no idea what s the correct...

in both get the error:

"validate and intialize error: Operator BSSVSoperator.c:LOC_disc has a statenode traitClockRate.c:LOC_disc in its inputs that is missing from the state."

can anyone help me figuring out what this means?

thanks in advance,

sara

[James Dorey]

Hi Sara,

I just had the same problem in BEAST 2.6.3 and managed to fix it.

The problem for me was that "traitClockRate.c:LOC_disc" should have actually been "ucldMean.c:LOC_disc". Changing it to this might fix your problem. I also set a lower bound of zero under there states:

        <parameter id="ucldMean.c:Pops" spec="parameter.RealParameter" lower="0.0" name="stateNode">1.0</parameter>

I found this out because it was what should have been called under the trait set block's branch rate model: 

                <branchRateModel id="RelaxedClock.c:Pops" spec="beast.evolution.branchratemodel.UCRelaxedClockModel" clock.rate="@ucldMean.c:Pops" rateCategories="@rateCategories.c:Pops" tree="@Tree.t:Ambaps_spatial_method3_ambiguity">

I don't know if you are still having trouble with this problem, but hopefully this helps if you are.

James 

[E.Ggnn]

Hi,

I have the same error message in my analyses.

I tried changing the line .

<parameter id=' traitClockRate.c :habit' name='mu' spec='parameter.RealParameter'>

to

<parameter id='ucldMean.c:habit' name='mu' spec='parameter.RealParameter'>

But then I get the following error message:

IDs should be unique. Duplicate id 'ucldMean.c:habit' found

Error detected about here:
  <beast>
      <run id='mcmc' spec='beast.coupledMCMC.CoupledMCMC'>
          <operator id='BSSVSoperator.c:habit' spec='BitFlipBSSVSOperator'>
              <parameter id='ucldMean.c:habit' name='mu' spec='parameter.RealParameter'>

I think I need to modify something in the state section, but haven't been able to figure out how to do it properly. I've attached the xml file, if anyone would like to try it out.

https://drive.google.com/file/d/1WIHAUrmqerELyhG9CurDnLHtqyIm4cLB/view?usp=sharing

With best wishes,

Edeline

[E.Ggnn]

Hi all,

I managed to solve my problem: it seems like there was some issues with building the xml with the way I had loaded my alignment file and set up my partitions in Beast 2. While I had 10 genes in my partition, I decided to reload them as 3 separate partitions, rather than 10, and then linking the clock, site and tree models accordingly. The xml file then formatted correctly!

Best,

Edeline