BEAST Classic, discrete mugration: java.lang.Error: randomChoiceUnnormalized falls through -- negative components in input distribution?
Pavel Rinkman
Hello, everyone
Please, I am asking for help from all those who have successfully used discrete phylogeography with BEAST Classic! 🙏
I am trying to run the discrete mugration model with BEAST Classic package following the tutorial. I did the model construction in BEAUTi word-by-word but got strange results. Actually, the results, trees & logs turned out normal, according to the tutorial. However, I caught the following in stderr:
Warning: removing transform class beast.base.inference.operator.kernel.Transform$LogConstrainedSumTransform because it should have at least 2 dimensions, but it has 0 Turning on scaling to prevent numeric instability 1.0201 … 0 NaN 1 NaN 2 NaN 3 NaN 4 NaN java.lang.Error: randomChoiceUnnormalized falls through -- negative components in input distribution? Please report error to MarcYou may find all files here
I see such the first time and do not understand the reason. Moreover, I have compared my XML with the example from the package GitHub repository and did not detect the difference, of course, besides new BEAST 2.7 default operators. I tried to run the example, but it falls with:
Error 1017 parsing the xml input file Class could not be found. Did you mean beast.base.inference.MCMC? Perhaps a package required for this class is not installed? Error detected about here: <beast> <run id='mcmc' spec='MCMC'>May this error influence the results of real analyses? How to correct my XMLs to eradicate it? I feel that I am silly due to the fact that I could not successfully run a simple guide and do not understand the nature of these errors.
Best regards,
Pavel
Neo One
Sou Noguchi
Hello everyone,
I am also facing the same or a very similar error in BEAST 2 / BEAST Classic during discrete phylogeographic analysis:
java.lang.Error: randomChoiceUnormalized falls through -- negative components in input distribution?
In my case, the stack trace includes AncestralStateTreeLikelihood and TreeWithTraitLogger.toNewick, so the problem seems to occur when BEAST tries to write the trait-annotated tree with ancestral location states.
The log file and ordinary trees file seem to be generated normally until the crash, but the run stops when the trait tree is logged. I am not sure whether this is a problem with my XML/model setting or a numerical issue in TreeWithTraitLogger.
If anyone has found a solution or workaround, I would greatly appreciate your help.