StarBeast2: gene tree constrained by species tree_Could not find a proper state to initialise

[janex...@gmail.com]

Hi StarBeast users and developers:

         I'm going to reconstruct a gene tree constrained by species tree and secondary calibration on the root, in BEAST v2.4.4.

         As it's functional gene coding regions and evolved constrained by species divergence. I try to find a suitable tree to reconstruct ancestral states and do selection analysis. 

         The error is as following, I didn't not figure out why. I attached the xml files I've tried and their error log files. 

         It's really appreciated if any one can help. Thanks so much!

Jia

===============================================================================

StarBEAST2: using fullInit to initialize all trees.

StarBEAST2: using fullInit to initialize all trees.

StarBEAST2: using fullInit to initialize all trees.

StarBEAST2: using fullInit to initialize all trees.

StarBEAST2: using fullInit to initialize all trees.

StarBEAST2: using fullInit to initialize all trees.

         P(posterior) = -Infinity (was -Infinity)

P(speciescoalescent) = 7.414923482038691 (was 7.414923482038691)

P(geneTree.t:Subset1) = 0.0 (was 0.0)

P(prior) = -Infinity (was -Infinity)

P(BirthDeathModel.t:Species) = 83.1791616526217 (was 83.1791616526217)

P(ExtinctionFractionPrior.t:Species) = 0.0 (was 0.0)

P(MultiMonophyleticConstraint.t:Species) = -Infinity (was -Infinity)

P(branchRatesStdevPrior.c:Subset1) = -1.4673450979966658 (was -1.4673450979966658)

P(clockRatePrior.c:Subset1) = -1.0439385332046727 (was -1.0439385332046727)

P(GammaShapePrior.s:Subset3) = -1.0 (was -1.0)

P(GammaShapePrior.s:Subset4) = -1.0 (was -1.0)

P(KappaPrior.s:Subset1) = -1.8653600339742873 (was -1.8653600339742873)

P(KappaPrior.s:Subset2) = -1.8653600339742873 (was -1.8653600339742873)

P(netDiversificationRatePrior.t:Species) = -1.4620820845188824 (was -1.4620820845188824)

P(HyperPrior.hyperLogNormalDistributionModel-M-netDiversificationRatePrior.t:Species) = 0.0 (was 0.0)

P(HyperPrior.hyperLStarBEAST2: using fullInit to initialize all trees.

StarBEAST2: using fullInit to initialize all trees.

StarBEAST2: using fullInit to initialize all trees.

StarBEAST2: using fullInit to initialize all trees.

Start likelihood: -Infinity after 10 initialisation attempts

ogNormalDistributionModel-S-netDiversificationRatePrior.t:Species) = -0.22314355131420976 (was -0.22314355131420976)

P(popMeanPrior.Species) = -1.4620820845188824 (was -1.4620820845188824)

P(PropInvariantPrior.s:Subset1) = 0.0 (was 0.0)

P(PropInvariantPrior.s:Subset2) = 0.0 (was 0.0)

P(PropInvariantPrior.s:Subset3) = 0.0 (was 0.0)

P(PropInvariantPrior.s:Subset4) = 0.0 (was 0.0)

P(RateACPrior.s:Subset3) = -1.8862943611198906 (was -1.8862943611198906)

P(RateAGPrior.s:Subset3) = -3.1686658147294304 (was -3.1686658147294304)

P(RateATPrior.s:Subset3) = -3.184008455701433 (was -3.184008455701433)

P(RateCGPrior.s:Subset3) = -3.184008455701433 (was -3.184008455701433)

P(RateCTPrior.s:Subset3) = -3.1686658147294304 (was -3.1686658147294304)

P(RateGTPrior.s:Subset3) = -3.184008455701433 (was -3.184008455701433)

P(Paroid_Root1.prior) = NaN (was NaN)  **

P(likelihood) = NaN (was NaN)  **

P(treeLikelihood.Subset1) = NaN (was NaN)  **

P(treeLikelihood.Subset2) = NaN (was NaN)  **

P(treeLikelihood.Subset3) = NaN (was NaN)  **

P(treeLikelihood.Subset4) = NaN (was NaN)  **

java.lang.RuntimeException: Could not find a proper state to initialise. Perhaps try another seed.

at beast.core.MCMC.run(Unknown Source)

at beast.app.BeastMCMC.run(Unknown Source)

at beast.app.beastapp.BeastMain.<init>(Unknown Source)

at beast.app.beastapp.BeastMain.main(Unknown Source)

at beast.app.beastapp.BeastLauncher.main(Unknown Source)

Fatal exception: Could not find a proper state to initialise. Perhaps try another seed.

Fatal exception: Could not find a proper state to initialise. Perhaps try another seed.

BEAST has terminated with an error. Please select QUIT from the menu.

[Huw A. Ogilvie]

Hi,

StarBEAST2 is not currently compatible with multi-monophyletic constraints. To fix the species tree it's better (in the sense of more efficient) to remove all species tree changing operators and use a newick starting tree. To set a starting tree add newick="<newick string>" to the attributes of the SBI element. So

    <init id="SBI" spec="starbeast2.StarBeastInitializer" birthRate="@netDiversificationRate.t:Species" estimate="false" speciesTree="@Tree.t:Species">

Becomes

    <init id="SBI" spec="starbeast2.StarBeastInitializer" birthRate="@netDiversificationRate.t:Species" estimate="false" speciesTree="@Tree.t:Species" newick="(A,B),C;">

Do you want to fix the species tree including branch lengths, or just the topology?

[janex...@gmail.com]

Hi Huw,

 

        Thanks for reply. I just need to fix topology of main clades, because the gene trees are in discordance with species trees. 

        I also tried to fix the complete topology in a starting newick tree in the  prior option - add multi-monophyletic constraints, as the  xml files uploaded. The difference was the monophyletic root or one of the clade. 

        Did you mean this method of add starting tree is not the source of error?

        Thanks again!

Jia

在 2017年3月15日星期三 UTC+8下午4:34:38，Huw A. Ogilvie写道：

[Huw A. Ogilvie]

Hi Jia,

The newick starting tree should be added to the SBI element, and a multi-monophyletic constraint should not be used with *BEAST/StarBEAST2. You can fix the entire species topology by removing the operators that change the topology of the species tree:

Reheight.t:Species
SubtreeSlide.t:Species
Narrow.t:Species
Wide.t:Species
WilsonBalding.t:Species

Or you can fix a few particular clades by adding them as monophyletic MRCA priors.

[Luiz JQ]

Dear Huw,

I was wondering if one of these operators can be kept in order to allow the analysis to play at least with the branch lengths. So basically, I would like to fix the topology but not the branch lengths. 

(I'm trying to construct a time-tree, but I want to use a fixed topology. As I just added a newick tree with equal branch lengths, my output tree shows exactly the same branch lengths, which is a little bizarre.)

Thank you very much.

Luiz

[Huw A. Ogilvie]

Hi Luiz,

The TreeScaler and TreeRootScaler elements will explore the space of branch lengths (and for StarBEAST2 the CoordinatedUniform and CoordinatedExponential operators do as well), so as long as you leave those elements in branch lengths will still be estimated. If you did leave those in, could you post the acceptance ratios of the operators? This is the output to the console after the chain has finished showing a bunch of counts for each operator.

- Huw

[Sophia Orzechowski]

Dear Huw,

1. To specify a starting species tree in a Starbeast2 analysis, what is the difference between the method you describe in this thread, and the method in the blog written by your colleague, Remco, https://www.beast2.org/2014/07/28/all-about-starting-trees? Remco suggests that with a Starbeast2 analysis, one should move the id="SBI" element above the run id="mcmc" element and follow the directions for *BEAST, which are quite involved compared to your method of adding a newick tree to the id="SBI" element.  

2. Secondly, to fix the topology of the specified species tree, it appears the method for *BEAST and Starbeast2 diverges. Why does Starbeast2 have 5 operators which control the topology of the species tree,  while *BEAST has only a single "Reheight.t:Species" operator one needs to set to zero to fix the topology?  

Many thanks for your time! 

All my best,

Sophie