Problem with output of Internode certainty analysis (-f i) in RAxML 8.2

[Kevin Kocot]

Hello,

I have been trying to use raxmlHPC-PTHREADS-8.2.0 to calculate internode certainty on a tree generated from a RAxML analysis of a supermatrix of 257 genes. The command I'm using is:
raxmlHPC-PTHREADS-8.2.0 -f i -T 7 -n IC_test -z trees.tre -t RAxML_bipartitions.FcC_smatrix.phy.out -m PROTGAMMAAUTO

where
FcC_smatrix.phy.out is the concatenated supermatrix used to generate RAxML_bipartitions.FcC_smatrix.phy.out generated using RAxML with -f a.
trees.tre is a file containing 257 single-gene trees (RAxML_bestTree* files) also generated using RAxML with -f a. Note that all taxa are not sampled for all genes.

The problem I'm having is that the output files RAxML_Corrected_Lossless_IC_Score_BranchLabels.IC_test and RAxML_Corrected_Probabilistic_IC_Score_BranchLabels.IC_test will not open in Figtree or MEGA and all branches appear to have the same length (0.90000000000000002220). I assume I'm doing something wrong but I don't know what it is. Any advice would be greatly appreciated.

Best,
Kevin

[Kassian Kobert]

Hello Kevin,

I am afraid the format of the output is not compatible with the tree viewing software at the moment.

If you manually inspect the RAxML_Corrected_Lossless_IC_Score_BranchLabels.IC_test and RAxML_Corrected_Probabilistic_IC_Score_BranchLabels.IC_test files with any standard text editor, you should see the IC/ICA values.

Best,

Kassian

[Karen]

Hi Kevin,

taken aside that Figtree at least has some bugs, try Dendroscope, TreeGraph or NJplot or Seaview (just a suggestion)

Best, karen

[Alexandros Stamatakis]

Thanks Karen,

Dendroscope works, you will need to enable the branch label viewing option.

Alexis

> --
> You received this message because you are subscribed to the Google
> Groups "raxml" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to raxml+un...@googlegroups.com
> <mailto:raxml+un...@googlegroups.com>.
> For more options, visit https://groups.google.com/d/optout.

--
Alexandros (Alexis) Stamatakis

Research Group Leader, Heidelberg Institute for Theoretical Studies
Full Professor, Dept. of Informatics, Karlsruhe Institute of Technology
Adjunct Professor, Dept. of Ecology and Evolutionary Biology, University
of Arizona at Tucson

www.exelixis-lab.org

[kmk...@gmail.com]

Thank you all for the assistance!

Best wishes,

Kevin

--
You received this message because you are subscribed to a topic in the Google Groups "raxml" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/raxml/TOItVeXYHQ0/unsubscribe.
To unsubscribe from this group and all its topics, send an email to raxml+un...@googlegroups.com.

[msp...@gmail.com]

Dear all,

I have the same problem: all branch lengths are set to "0.90000000000000002220" , but it is not just an issue of compatibility with the tree viewer.
When I annotate a reference tree with IC support from a collection of partial gene trees like this:

raxmlHPC-AVX -m GTRCAT -f i -t reference_tree.tre -z partial_trees.tre -n IC_run
Version: 8.2.4
Reference tree was made with examl, if that matters.

The output tree does contain the topology and IC scores etc, but not the original branch lengths of the reference tree, instead all branches are set to length "0.90000000000000002220" as Kevin observed! (checked this on the newick file, not a viewer).
How can I recover an output with both the reference branch lengths and the IC scores?

Regarding the viewer Issue: Here is some messy sed / perl to convert to Figtree - readable newick format, including negative scores:
cat RAxML_Corrected_Probabilistic_IC_Score_BranchLabels.IC_run | sed 's/,.....]/]/g' | sed 's/,-.....]/]/g' | perl -p -i -e 's/:(\d+\.\d+)\[(.....)]/$2:$1/g' | perl -p -i -e 's/:(\d+\.\d+)\[(......)]/$2:$1/g' > mytree_figtree_readable.txt

regards,
Mathias

[Alexandros Stamatakis]

Hi Mathias,

Thanks for pointing this out, I will put this on my TODO list.

Alexis

[Jacek Kominek]

Hi Alexis,

I just wanted to +1 this issue, since it's still present in the latest RAxML (8.2.10 from github). When running "-f i" all branches in the output are 0.90...0022 (with many zerous in the middle). Interestingly enough, it's only present if the partial correction needs to be done - if all taxa are present in the gene trees passed via "-z" then the branch lengths are written out properly. I'm sure it's a fairly simple fix, but I assume you have more pressing matters at the moment, especially since this issue could easily be worked around by a custom Python/Perl script, so it's not a very big deal. Nevertheless, just wanted to bump this one, when you have time for it. Thanks!

-Jacek

[Alexandros Stamatakis]

thanks jacek,

can you please send me the dataset and command line where this occurs, I
will try to fix it if possible,

alexis

On 16.06.2017 16:37, Jacek Kominek wrote:
> Hi Alexis,
> I just wanted to +1 this issue, since it's still present in the latest
> RAxML (8.2.10 from github). When running "-f i" all branches in the
> output are 0.90...0022 (with many zerous in the middle). Interestingly
> enough, it's only present if the partial correction needs to be done -
> if all taxa are present in the gene trees passed via "-z" then the
> branch lengths are written out properly. I'm sure it's a fairly simple
> fix, but I assume you have more pressing matters at the moment,
> especially since this issue could easily be worked around by a custom
> Python/Perl script, so it's not a very big deal. Nevertheless, just
> wanted to bump this one, when you have time for it. Thanks!
>
> -Jacek
>
> W dniu poniedziałek, 21 listopada 2016 14:06:00 UTC-6 użytkownik Alexis
> napisał:
>
> Hi Mathias,
>
> Thanks for pointing this out, I will put this on my TODO list.
>
> Alexis
>

> > an email to raxml+un...@googlegroups.com <javascript:>
> > <mailto:raxml+un...@googlegroups.com <javascript:>>.

> > For more options, visit https://groups.google.com/d/optout

> <https://groups.google.com/d/optout>.

>
> --
> Alexandros (Alexis) Stamatakis
>
> Research Group Leader, Heidelberg Institute for Theoretical Studies
> Full Professor, Dept. of Informatics, Karlsruhe Institute of Technology
> Adjunct Professor, Dept. of Ecology and Evolutionary Biology,
> University
> of Arizona at Tucson
>

> www.exelixis-lab.org <http://www.exelixis-lab.org>

[Jacek Kominek]

Sure thing, sent a sample dataet directly to your email account. Hopefully it helps.

-J

[Alexandros Stamatakis]

this should be fixed now, please check, just released version 8.2.11

https://github.com/stamatak/standard-RAxML/releases/tag/v8.2.11

alexis

www.exelixis-lab.org

[Jacek Kominek]

Just tested, works like a charm, thanks a bunch! :)

-J

[Perrine Mandon]

I got kind of a similar issue using the command :

raxmlHPC -T 4 -f a -p 12345 -x 12345 -m GTRGAMMAI -#500 -s DataB2.nex -n DataB2-ML.tre -q Concat_partitions.txt

(on RAxML version 8.2.9 )
and my branch lengths were always 0.10536051565782628137

I finally unerstood why :
It is because my 2nd positions of H3 codon are exactly the same for all my taxa.
As I partitionned my fasta file by gene and codon, I gave a partition with no variations.
If I just keep positions 1 and 3 of this gene, it seems to work well.