Combined morphological and molecular data

[Bernardo Santos]

Hi all,

I'm trying to run a combined morphology+molecular analysis. At first I got the error message of "numerical underflow" that calls for using the -safe option. I rerun with the -safe option but now I get the following error:

WARNING: Numerical underflow for lh-branch

ERROR: partitionmodelplen.cpp:142: virtual double PartitionModelPlen::optimizeParameters(int, bool, double, double): Assertion `cur_lh > tree_lh - 1.0 && "partition rate opt reduces LnL"' failed.

ERROR: STACK TRACE FOR DEBUGGING:

ERROR:

ERROR: *** IQ-TREE CRASHES WITH SIGNAL ABORTED

ERROR: *** For bug report please send to developers:

ERROR: ***    Log file: partitions.nex.log

ERROR: ***    Alignment files (if possible)

I called IQtree with the following command: iqtree -spp partitions.nex -m TESTNEWMERGE -bb 1000 -safe

My partition, log and alignment files are here: https://www.dropbox.com/s/etm6by8mdb0ewv8/Santos_IQtree_combined_alignments.zip.zip?dl=0

Would really appreciate any insight here! Thanks!

[Eric Smidt]

I use a partitioned nexus file

#nexus
begin sets;
 charset part1 = mol117.nexus: 1-769;
 charset part7 = morfo117p.nexus: 1-39;
 charpartition mine = TIM3e+I+G4:part1, MK+ASC:part2;
end;

and use this command

bin\iqtree -sp partitionedctst.nex -alrt 1000 -bb 1000 -bnni

works fine

Eric de Camargo Smidt
Prof. Associado II
Laboratório de Sistemática e Ecologia Molecular de Plantas
Departamento de Botânica, sala 446
Setor de Ciências Biológicas Centro Politécnico - UFPR
Caixa Postal 19031 Cep 81531-990
Curitiba, PR - Brasil
e-mail: ecs...@ufpr.br
http://orcid.org/0000-0002-1177-1682

https://publons.com/researcher/2905687/eric-smidt/
http://lattes.cnpq.br/1539716845687200

https://sites.google.com/view/lsemp-ufpr

--
You received this message because you are subscribed to the Google Groups "IQ-TREE" group.
To unsubscribe from this group and stop receiving emails from it, send an email to iqtree+un...@googlegroups.com.
To post to this group, send email to iqt...@googlegroups.com.
Visit this group at https://groups.google.com/group/iqtree.
To view this discussion on the web visit https://groups.google.com/d/msgid/iqtree/bba6d73d-3084-4e2d-a8e5-65202a381ca5%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Minh Bui]

Hi Bernardo Santos,

Thanks for the report. 

I confirm that there is a bug with -spp option and mixed data type. We will fix in the next release. So please wait until I send further notice.

However, there is also a problem with your data: the DNA and morphology partitions have no taxa in common. So please check it out.

Cheers,

Minh

To view this discussion on the web visit https://groups.google.com/d/msgid/iqtree/CAN3iz%2BkJW-5B_KMgR1CK4s3vpjVFOseoXuT6WBQpiLoByyMHyg%40mail.gmail.com.

[Minh Bui]

Sorry, to be more precise: this bug happens in the model selection step with these two options. Otherwise, iqtree works fine as Eric pointed out below.

So as a quick workaround, you can do the following: perform model selection separately for DNA and morphology data. Then look at the two partition files, combine them into one file by merging the two charpartition commands. And now you can run iqtree on this combined partition file.

Cheers,

Minh