StarBEAST2 "no lineages for species taxon X" error

[Andrés Parada]

Dear beast users,

I got this error analyzing 20 loci (all of them with the same number of taxa, fully "occupied") with StarBEAST2 and 2.4.4 version following the September 28, 2016 tutorial. Using Ubuntu 16.04.1 java version "1.8.0_91"

10814 is the first gene in the analysis. specie1 is the first taxon too. How could I fix this? Thanks for your help.

StarBEAST2: using fullInit to initialize all trees.
Fatal exception: Gene tree Tree.t:10814 has no lineages for species taxon specie1
java.lang.RuntimeException: Gene tree Tree.t:10814 has no lineages for species taxon specie1
 at starbeast2.StarBeastInitializer.fullInit(StarBeastInitializer.java:232)
 at starbeast2.StarBeastInitializer.initStateNodes(StarBeastInitializer.java:101)
 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: Gene tree Tree.t:10814 has no lineages for species taxon specie1
java.lang.RuntimeException: An error was encounted. Terminating BEAST
 at beast.app.util.ErrorLogHandler.publish(Unknown Source)
 at java.util.logging.Logger.log(Logger.java:738)
 at java.util.logging.Logger.doLog(Logger.java:765)
 at java.util.logging.Logger.log(Logger.java:788)
 at java.util.logging.Logger.severe(Logger.java:1463)
 at beast.app.beastapp.BeastMain.<init>(Unknown Source)
 at beast.app.beastapp.BeastMain.main(Unknown Source)
 at beast.app.beastapp.BeastLauncher.main(Unknown Source)

[Andrés Parada]

StarBEAST2 version according to the app manager is 0.13.1

[Huw A. Ogilvie]

Hi Andrés, could you post a screenshot of your Taxon Sets tab in BEAUti, and/or the XML file used for the analysis?

[Andrés Parada]

Hi Huw, I just saw the reply, I am at home right now but the scheme is just like below (29 species), I have one single sequence per gene per species. I have full occupancy for all genes/species

specie1 specie1

specie2 specie2

Thanks

--
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.

--

Postdoctoral Associate

Instituto de Cs Ambientales y Evolutivas

Universidad Austral de Chile

Campus Isla Teja s/n

Valdivia, Chile

[Huw A. Ogilvie]

Andrés, I think I need to look at the XML or source nexus files to diagnose what's going on

[Huw Ogilvie]

Hi Andrés,

The names of sampled individuals must be different from the names of species.

So you can use a naming scheme like this: SpeciesName SpeciesName1

But you can't use a naming scheme like this: SpeciesName SpeciesName

Or one like this: SpeciesName1 SpeciesName1

So either change the name of each sequence in the FASTA/Nexus files to be different from the species names, or change the species names in the Taxon Sets tab to be different from the individual names.

Regarding the number of genes, the StarBEAST2 paper has a number of informative analyses. For a dataset with 19 sampled individuals and 100 genes, SpeciesTreeUCLN analyses ran at about 1 ESS per hour, and would normally finish within 1 month. For a data set with 42 sampled individuals and 52 genes, SpeciesTreeUCLN ran at about 5 ESS per hour, and would normally finish within 1 week. Given your data set has 29 sampled individuals, I would suggest using the sample data set as for "classic" BEAST analysis, although this may take a few days or a few weeks to achieve a sufficient ESS for parameters of interest. It is difficult to pin down an exact time frame because the speed of Bayesian phylogenetics is highly depending on the particular data set (see Ogilvie 2016 in Systematic Biology).

On 2 February 2017 at 06:49, Andrés Parada <andre...@gmail.com> wrote:

Hi Huw, 

Please find attached the xml file, 

Can I ask you if 20 are too much genes or how long it might last the run? I have a total ~300 genes. I picked an arbitrary number among the 68 fast evolving genes (or with less invariant sites)  to do this particular analysis after doing it with 68 genes and a "classic" beast analysis  (using ONE clock, and under 2.4.3)

For this analysis I used the template SpeciesTreeUCLN and this file (also attached) names for guessing the taxon sets.

Thanks for your help

Cheers

[Huw Ogilvie]

My apologies, using the SAME data set as for the "classic" BEAST analysis

[Andrés Parada]

I will fix the names file and   try a bigger dataset then.

Thank you very much for your help

[fro...@gmail.com]

Hi, my name's Aaron, just joined. I'm using *BEAST2 2.5 on a dataset of 7 loci and about 50 specimens, with different taxon names from the specimen names, but I'm getting the same error as Andres. Is there an issue if not every taxon is represented in every single locus?

[Huw A. Ogilvie]

Hi Aaron,

If you are using the StarBEAST2 template, there should be no need to have every taxon represented in the molecular data. Could you upload your XML file somewhere or send it to me so I can take a look at what is causing the error?

- Huw

[Aaron Fronk]

Sure! Can you get it if I upload it here, so?

To unsubscribe from this group and stop receiving emails from it, send an email to beast-users+unsubscribe@googlegroups.com.

[Aaron Fronk]

Sorry, sent the wrong version. That last one had been modified; this is the one I tried running last.

[Huw A. Ogilvie]

Hi Aaron,

That XML was created using the old "StarBEAST" tempate, which does not work unless there is at least one sequence for each species for each locus. If you install StarBEAST2 in the package manager and use the StarBEAST2 template, then that should fix your issue.

- Huw