Different trees in a posterior sample according to sumtrees.py.

[Bastien Boussau]

Hi,

I used sumtrees.py to extract the distinct trees and their frequencies in a sample of trees. Sumtrees told me there were 4 different topologies. I got a weird result however in that topologies 1 and 3 were the same, as shown below. Rooted Phylip consense tells me that there are indeed 4 different trees, that differ only by the position of their root, so I'm thinking there may be a problem with the output, at least for topology 3. I'm using a git version of sumtrees I downloaded on March 1st.

Here is the command I used:
python ~/Downloads/dendropy/DendroPy/scripts/sumtrees/sumtrees.py -b 5000 --ultrametric --ultrametricity-precision 0.1  --support-as-labels --output=consensusAllTreesB5000.tre outputTreesOMPAll.txt --trprobs=treesWithFrequency.tre

and the content of the file treesWithFrequency.tre:
more treesWithFrequency.tre
#NEXUS



BEGIN TAXA;

    DIMENSIONS NTAX=36;
    TAXLABELS
        ANASP
        ANAVT
        NOSP7
        TRIEI
        CYAA5
        CYAP8
        CYAP7
        MICAN
        SYNY3
        SYNP2
        THEEB
        CYAP4
        ACAM1
        SYNR3
        PROMM
        PROM3
        PRMAR1
        PROM4
        PROM1
        PROMT
        PROM9
        PROM2
        PROM0
        PROMS
        PROM5
        PROMP
        SYNPW
        SYNS3
        SYNPX
        SYNS9
        SYNSC
        SYNE7
        SYNP6
        SYNJA
        SYNJB
        GLVIO1
  ;
END;

BEGIN TREES;

    TREE Tree1 = [&count=21795,probability=0.853467517719,cumulative_probability=0.853467517719](ANASP,ANAVT,(NOSP7,(TRIEI,((SYNP2,(SYNY3,((CYAA5,CYAP8),(CYAP7,MICAN)))),((ACAM1,(THEEB,CYAP4)),(((SYNR3,(((PROMM,PROM3),(((PROM1,PROMT),((PROM5,PROMP),(PROM9,(PROM2,(PROM0,PROMS))))),(PRMAR1,PROM4))),((SYNPW,SYNS3),(SYNPX,(SYNS9,SYNSC))))),(SYNE7,SYNP6)),(GLVIO1,(SYNJA,SYNJB))))))));

    TREE Tree2 = [&count=2778,probability=0.108783333986,cumulative_probability=0.962250851705](ANASP,ANAVT,(NOSP7,(TRIEI,(((ACAM1,(THEEB,CYAP4)),(((SYNE7,SYNP6),(SYNR3,(((SYNPW,SYNS3),(SYNPX,(SYNS9,SYNSC))),((PROMM,PROM3),(((PROM1,PROMT),((PROM5,PROMP),(PROM9,(PROM2,(PROM0,PROMS))))),(PRMAR1,PROM4)))))),(GLVIO1,(SYNJA,SYNJB)))),(SYNP2,(SYNY3,((CYAA5,CYAP8),(CYAP7,MICAN))))))));

    TREE Tree3 = [&count=708,probability=0.0277244782081,cumulative_probability=0.989975329913](ANASP,ANAVT,(NOSP7,(TRIEI,((SYNP2,(SYNY3,((CYAA5,CYAP8),(CYAP7,MICAN)))),((ACAM1,(THEEB,CYAP4)),(((SYNR3,(((PROMM,PROM3),(((PROM1,PROMT),((PROM5,PROMP),(PROM9,(PROM2,(PROM0,PROMS))))),(PRMAR1,PROM4))),((SYNPW,SYNS3),(SYNPX,(SYNS9,SYNSC))))),(SYNE7,SYNP6)),(GLVIO1,(SYNJA,SYNJB))))))));

    TREE Tree4 = [&count=256,probability=0.0100246700865,cumulative_probability=1.0](ANASP,ANAVT,(NOSP7,(TRIEI,((SYNP2,(SYNY3,((CYAA5,CYAP8),(CYAP7,MICAN)))),((ACAM1,(THEEB,CYAP4)),((GLVIO1,(SYNJA,SYNJB)),((SYNR3,(((PROMM,PROM3),(((PROM1,PROMT),((PROM5,PROMP),(PROM9,(PROM2,(PROM0,PROMS))))),(PRMAR1,PROM4))),((SYNPW,SYNS3),(SYNPX,(SYNS9,SYNSC))))),(SYNE7,SYNP6))))))));

END;

I can send the tree file to someone who would want to repeat the experiment...

Best,

Bastien.

[Bastien Boussau]

Jeremy Brown pointed out to me that the trees in the sumtrees output were unrooted, which explains why I saw no difference between trees 1 and 3. If I add the option "--rooted" to the command line, the trees are still unrooted. Is there a way to get rooted trees in the trprobs file?
Thanks!

[Jeet Sukumaran]

Hi Bastian,

I will look into this: could you please submit an issue on the GitHub
site so it does not slip through my porous memory? If it is not too
large, it would be good to attach the original tree file.

Is it possible for you to check whether the clade support calculated
by SumTrees is correct, by, for e.g. running the trees through raxml
or PAUP*?

Do the original trees explicitly declare their rooting using "[&R]"
comment? The "--rooted" option should override this in any case, but
if there is a bug in propagating this option down to the topology
counter, maybe adding "[&R]" would provide a kludge work around? Of
course, it may already be present or may not help ...

-- jeet

[Bastien Boussau]

Hi Jeet,

Sorry for this very long silence. I downloaded the latest version of
sumtrees and I still have the problem I was reporting months ago.
I submitted an issue on the github repository.

Thanks for your help!

Best,

Bastien.