trouble with initial beast run

[Chris Law]

Hello,

I am trying to run my xml file in Beast 2.1.3 but I receive this error: 

java.lang.Exception: 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 sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)

at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)

at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)

at java.lang.reflect.Method.invoke(Method.java:597)

at apple.launcher.LaunchRunner.run(LaunchRunner.java:116)

at apple.launcher.LaunchRunner.callMain(LaunchRunner.java:51)

at apple.launcher.JavaApplicationLauncher.launch(JavaApplicationLauncher.java:52)

Anybody have any suggestions? I've attached my xml file.

Thanks for the help.

[Remco Bouckaert]

Hi Chris,

It looks like the files sets proportion invariant parameters to 1.0 initially. It is expected to be between zero and one, and setting it to one initially results in infinite likelihoods (this should be picked up, but is not at the moment).

To make the file run, set all proportion invariant entries to something in between 0 and 1.

By the way, the site model for the ADORA3 partition does not use gamma heterogeneity, while the other partitions do.

Cheers,

Remco

[Kasia Mikolajczak]

Hello,

I have the same error popping up for the first exploratory BEAST run for my data.

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

I use the BEAST v 2.1.3 on OS X v. 10.9.3

I’ve read the advice given on the issue before, but although I’ve set the proportion invariant sites to 0.5 in all alignments for now, the problem persists. I am wondering if perhaps it may have something to do with my nexus files: in two of them I had to add an one unknown sequence (Onit0383) because otherwise the taxon would not be represented in these two alignments at all. Any advice will be greatly appreciated.

Best wishes

Kasia

[Remco Bouckaert]

Hello Kasia,

Is it possible that one of the species in this *BEAST analysis is not represented by all loci If so, you need to either add more sequences so all species are represented by at least one sequence, or remove the alignment altogether.

Cheers,

Remco

[Kasia Mikolajczak]

Hi Remco,

Thank you for replying. It's true I don't have a single sequence for one of the species in two of the alignments, but I was under the impression that including a dummy sequence filled with N or ? should suffice (at least this guy here says so: http://jimmyodonnell.wordpress.com/category/science/)

However, even the removal of the two alignments from the analysis doesn't resolve the problem, *BEAST run stops at exactly same stage, so it must be something else. I just don't know what.

Best wishes,

Kasia

[Chris Law]

Hi Remco,

Thanks so much for the advice. I got it running.

Cheers,

[Kritika Garg]

Hi Remco,
                 I am getting the same error with star beast. I have at least one sequence per species. Please let me know what can I do.

Thank you

Regards
Kritika

Here is the error
Start likelihood: NaN
P(posterior) = NaN (was NaN)
    P(speciescoalescent) = NaN (was NaN)
        P(SpeciesTreePopSize.Species) = NaN (was NaN)
        P(treePrior.t:cytb) = NaN (was NaN)
        P(treePrior.t:dloop) = NaN (was NaN)
        P(treePrior.t:prkc) = NaN (was NaN)
        P(treePrior.t:tg) = NaN (was NaN)
    P(prior) = NaN (was NaN)
        P(YuleModel.t:Species) = NaN (was NaN)
        P(YuleBirthRatePrior.t:Species) = -5.362240542149371 (was NaN)
        P(ClockPrior.c:dloop) = 0.0 (was NaN)
        P(ClockPrior.c:prkc) = 0.0 (was NaN)
        P(ClockPrior.c:tg) = 0.0 (was NaN)
        P(GammaShapePrior.s:hky+g) = -1.0 (was NaN)
        P(KappaPrior.s:cytb) = -1.8653600339742873 (was NaN)
        P(KappaPrior.s:hky+g) = -1.8653600339742873 (was NaN)
        P(popMean.prior) = 0.0 (was NaN)
        P(ucldStdevPrior.c:clock) = -0.40143772133305716 (was NaN)
    P(likelihood) = NaN (was NaN)
        P(treeLikelihood.cytb) = NaN (was NaN)
        P(treeLikelihood.dloop) = NaN (was NaN)
        P(treeLikelihood.prkc) = NaN (was NaN)
        P(treeLikelihood.tg) = NaN (was NaN)

java.lang.Exception: 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 sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)
    at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
    at java.lang.reflect.Method.invoke(Method.java:597)
    at apple.launcher.LaunchRunner.run(LaunchRunner.java:116)
    at apple.launcher.LaunchRunner.callMain(LaunchRunner.java:51)
    at apple.launcher.JavaApplicationLauncher.launch(JavaApplicationLauncher.java:52)

[Kritika Garg]

Hi Remco,
                  I have the same error. I have attached the xml file. I have one sequence per gene for each species. Can you please have a look at the xml file and give me your suggestion.

Thank you

Regards
Kritika

Start likelihood: NaN
P(posterior) = NaN (was NaN)
    P(speciescoalescent) = NaN (was NaN)
        P(SpeciesTreePopSize.Species) = NaN (was NaN)
        P(treePrior.t:cytb) = NaN (was NaN)
        P(treePrior.t:dloop) = NaN (was NaN)
        P(treePrior.t:prkc) = NaN (was NaN)
        P(treePrior.t:tg) = NaN (was NaN)
    P(prior) = NaN (was NaN)
        P(YuleModel.t:Species) = NaN (was NaN)
        P(YuleBirthRatePrior.t:Species) = -5.362240542149371 (was NaN)
        P(ClockPrior.c:dloop) = 0.0 (was NaN)
        P(ClockPrior.c:prkc) = 0.0 (was NaN)
        P(ClockPrior.c:tg) = 0.0 (was NaN)
        P(GammaShapePrior.s:hky+g) = -1.0 (was NaN)
        P(KappaPrior.s:cytb) = -1.8653600339742873 (was NaN)
        P(KappaPrior.s:hky+g) = -1.8653600339742873 (was NaN)
        P(popMean.prior) = 0.0 (was NaN)
        P(ucldStdevPrior.c:clock) = -0.40143772133305716 (was NaN)
    P(likelihood) = NaN (was NaN)
        P(treeLikelihood.cytb) = NaN (was NaN)
        P(treeLikelihood.dloop) = NaN (was NaN)
        P(treeLikelihood.prkc) = NaN (was NaN)
        P(treeLikelihood.tg) = NaN (was NaN)

java.lang.Exception: 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 sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)
    at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
    at java.lang.reflect.Method.invoke(Method.java:597)
    at apple.launcher.LaunchRunner.run(LaunchRunner.java:116)
    at apple.launcher.LaunchRunner.callMain(LaunchRunner.java:51)
    at apple.launcher.JavaApplicationLauncher.launch(JavaApplicationLauncher.java:52)

[Remco Bouckaert]

Hi Kritika,

It looks like there are some gene trees that do not have any genes for every of the species taxa. For example tg has no sequences for MA. *BEAST only runs when each species has some sequences for each species. If you do not have this data, you can either remove the gene from the analysis or add a dummy sequence with all data missing for these cases.

Cheers,

Remco

--
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 http://groups.google.com/group/beast-users.
For more options, visit https://groups.google.com/d/optout.
<test.xml>

[Marcial Escudero]

Dear Remco Bouckaert,

I am having the same problem. I have checked several of your suggestions but any of them seem to work for me.

I was working with a dataset of 4 genes (scentaurium.xml) but for one the genes I had not all the sequences (I added dummy sequences with all data missing for these as you suggested).
I have also decided to excluded one of the genes (scentaurium2.xml) but I am getting the same problem.

I have tried to change the distribution in the clock from normal to lognormal. Or changing some parameters (following your suggestions) but nothing seems to work.

Thanks!

Marcial.


See the message:

Start likelihood: -Infinity after 10 initialisation attempts
P(posterior) = -Infinity (was NaN)
    P(speciescoalescent) = 956.3427921833753 (was NaN)
        P(SpeciesTreePopSize.Species) = 264.48901083565744 (was NaN)
        P(treePrior.t:cent7c) = 207.6614741663799 (was NaN)
        P(treePrior.t:cent8) = 199.47668217691523 (was NaN)
        P(treePrior.t:centplastidc) = 284.7156250044228 (was NaN)
    P(prior) = -Infinity (was NaN)
        P(YuleModel.t:Species) = 95.6006770539091 (was NaN)
        P(YuleBirthRatePrior.t:Species) = -3.9660643081713065 (was NaN)
        P(GammaShapePrior.s:cent7c) = -1.0 (was NaN)
        P(GammaShapePrior.s:cent8) = -1.0 (was NaN)
        P(GammaShapePrior.s:centplastidc) = -1.0 (was NaN)
        P(popMean.prior) = 5.947010904276155 (was NaN)
        P(RateACPrior.s:cent7c) = -3.184008455701433 (was NaN)
        P(RateACPrior.s:cent8) = -3.184008455701433 (was NaN)
        P(RateACPrior.s:centplastidc) = -3.184008455701433 (was NaN)
        P(RateAGPrior.s:cent7c) = -3.1686658147294304 (was NaN)
        P(RateAGPrior.s:cent8) = -3.1686658147294304 (was NaN)
        P(RateAGPrior.s:centplastidc) = -3.1686658147294304 (was NaN)
        P(RateATPrior.s:cent7c) = -3.184008455701433 (was NaN)
        P(RateATPrior.s:cent8) = -3.184008455701433 (was NaN)
        P(RateATPrior.s:centplastidc) = -3.184008455701433 (was NaN)
        P(RateCGPrior.s:cent7c) = -3.184008455701433 (was NaN)
        P(RateCGPrior.s:cent8) = -3.184008455701433 (was NaN)
        P(RateCGPrior.s:centplastidc) = -3.184008455701433 (was NaN)
        P(RateGTPrior.s:cent7c) = -3.184008455701433 (was NaN)
        P(RateGTPrior.s:cent8) = -3.184008455701433 (was NaN)
        P(RateGTPrior.s:centplastidc) = -3.184008455701433 (was NaN)
        P(MeanRatePrior.c:cent7c) = 0.0 (was NaN)
        P(MeanRatePrior.c:cent8) = 0.0 (was NaN)
        P(MeanRatePrior.c:centplastidc) = 0.0 (was NaN)
        P(ucldStdevPrior.c:cent7c) = 0.819364145120802 (was NaN)
        P(ucldStdevPrior.c:cent8) = 0.819364145120802 (was NaN)
        P(ucldStdevPrior.c:centplastidc) = 0.819364145120802 (was NaN)
        P(CentauriumETS.prior) = -Infinity (was NaN)
        P(CentauriumITS.prior) = NaN (was NaN)
        P(Centauriumplastid.prior) = NaN (was NaN)
        P(SECETS.prior) = NaN (was NaN)
        P(SECITS.prior) = NaN (was NaN)
        P(SECplastid.prior) = NaN (was NaN)

    P(likelihood) = NaN (was NaN)

        P(treeLikelihood.centplastidc) = NaN (was NaN)
        P(treeLikelihood.cent7c) = NaN (was NaN)
        P(treeLikelihood.cent8) = NaN (was NaN)

java.lang.Exception: 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)
amesclir@rsantos:~/local/bin/beast2.3.1/beast/bin$

[chandith1]

Hi Remco
I also have the same problem when trying to run BDSKY model

Start likelihood: -Infinity after 10 initialisation attempts

P(posterior) = -Infinity (was -Infinity)
    P(prior) = -Infinity (was -Infinity)
        P(BirthDeathSkySerial.t:TEST1) = -2850.6703955878056 (was -2850.6703955878056)
        P(becomeUninfectiousRatePrior.t:TEST1) = -0.9189385332046727 (was -0.9189385332046727)
        P(GammaShapePrior.s:TEST1) = -1.0 (was -1.0)
        P(originPrior.t:TEST1) = 0.0 (was 0.0)
        P(HyperPrior.hyperBeta-alpha-samplingProportionPrior.t:TEST1) = 0.0 (was 0.0)
        P(HyperPrior.hyperBeta-beta-samplingProportionPrior.t:TEST1) = 0.0 (was 0.0)
        P(HyperPrior.hyperLogNormalDistributionModel-M-becomeUninfectiousRatePrior.t:TEST1) = -Infinity (was -Infinity)
        P(HyperPrior.hyperLogNormalDistributionModel-M-reproductiveNumberPrior.t:TEST1) = -Infinity (was -Infinity)
        P(HyperPrior.hyperLogNormalDistributionModel-S-becomeUninfectiousRatePrior.t:TEST1) = 0.0 (was 0.0)
        P(HyperPrior.hyperLogNormalDistributionModel-S-reproductiveNumberPrior.t:TEST1) = 0.0 (was 0.0)
        P(RateACPrior.s:TEST1) = -3.184008455701433 (was -3.184008455701433)
        P(RateAGPrior.s:TEST1) = -3.1686658147294304 (was -3.1686658147294304)
        P(RateATPrior.s:TEST1) = -3.184008455701433 (was -3.184008455701433)
        P(RateCGPrior.s:TEST1) = -3.184008455701433 (was -3.184008455701433)
        P(RateCTPrior.s:TEST1) = -3.1686658147294304 (was -3.1686658147294304)
        P(RateGTPrior.s:TEST1) = -3.184008455701433 (was -3.184008455701433)
        P(reproductiveNumberPrior.t:TEST1) = -18.523122207237183 (was -18.523122207237183)
        P(samplingProportionPrior.t:TEST1) = 8.881784197001248E-16 (was 8.881784197001248E-16)
        P(MeanRatePrior.c:TEST1) = 0.0 (was 0.0)
        P(ucldStdevPrior.c:TEST1) = 0.819364145120802 (was 0.819364145120802)
    P(likelihood) = NaN (was NaN)  **
        P(treeLikelihood.TEST1) = NaN (was NaN)  **
    P(fossilCalibrations) = 0.0 (was NaN)  **
Fatal exception: Could not find a proper state to initialise. Perhaps try another seed.
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.

Pls assist