Exception in thread "main" java.lang.StackOverflowError - STARBEAST error

[aloliveira]

Hello all,

I am having problems with beast2. I edited and compared several .xml files and I think the problem might be in the   <taxonset id="taxonsuperset" spec="TaxonSet"> specifications. At first beauti2 did not created all the taxon idref for my sequences (bug maybe??), problem that I fixed manually (or I think that I fixed). Since then, I have been facing java.lang.StackOverflowError errors, however that seems a bit improbable because I don't have too many taxa in my analysis (only 74). Does anyone have an idea of how can I solve this problem?

I am attaching the .xml file here, if someone is able to help me. This is the error message that I keep getting.

Exception in thread "main" java.lang.StackOverflowError
    at java.security.AccessController.doPrivileged(Native Method)
    at java.net.URLClassLoader.findClass(URLClassLoader.java:361)
    at java.lang.ClassLoader.loadClass(ClassLoader.java:424)
    at java.lang.ClassLoader.loadClass(ClassLoader.java:411)
    at sun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:338)
    at java.lang.ClassLoader.loadClass(ClassLoader.java:357)
    at java.lang.Class.forName0(Native Method)
    at java.lang.Class.forName(Class.java:264)
    at beast.util.XMLParser.resolveClass(Unknown Source)
    at beast.util.XMLParser.parseInputs(Unknown Source)
    at beast.util.XMLParser.createObject(Unknown Source)
    at beast.util.XMLParser.createObject(Unknown Source)
    at beast.util.XMLParser.parseInputs(Unknown Source)

All the best,
André

[Remco Bouckaert]

Hi André,

The main issue was that the species names are equal to the gene names. So, when BEAST was trying to find out what the tax onset of a species looked like, it found out it contained a taxon consisting of that same taxon set and went into a loop, only to quit when running out of stack space. To fix this, I renamed all species tree taxa by inserting a T in the name, and for every taxon “XYZ"replaced <taxon idref=“XYZ”/> with <taxon id=“XYZ” spec=“Taxon”/> to specify a new taxon. 

Two more minor things: The sequence 611_H_gut was removed, so I added it back in. It turned out that in the SiteModel for rps7_GENE, the shape parameter was set to 0, but (even though this value is ignored since there is no gamma rate heterogeneity for this site model) it should be positive, so I set it to 1.0.

With these changes, the file runs now.

Cheers,

Remco