Topological constraints and Consel tests

[Tracey Nowell]

Hi, I press 'send' with a great sense of trepidation! Please be gentle.

Can someone please tell me if the tests that I have done in Consel are valid/make any sense. I've attached outputs, they seem off the scale and it has been suggested to me that there is something wrong, although all sections of the analyses went swimmingly! Perhaps these values are valid and the constrained tree is wildly less likely than the other?

I got per site log likelihoods from comparison of two trees generated using RAxML, one with and one without a topological constraint (tree file attached).

I’ve looked through previous postings and thought that I may be experiencing the same problem as Gregory Mendez back in October 2015 – constrained group not strictly monophyletic (second tree in the file), but I’m not sure that it isn't (as Alexis says in the manual) that taxa not under a constraint are free to fall into any part of the tree. I have rooted the tree on the appropriate outgroup and it makes no difference. The two TracemXXXX taxa are the ones that I "forced" into the same clade as all other 'T" taxa, but following ML optimisation they seem to be part of a trichotomy.

For constraint RAxML was called as follows:

raxmlHPC -f a -x 12479 -p 14507 -# 1000 -m GTRGAMMA -q geneparts.txt -s kalan87.phy -g constraint.tre -n a5

I would really appreciate any advice on this.

Many thanks

Tracey

[Alexandros Stamatakis]

Hi Tracey,

> Can someone please tell me if the tests that I have done in Consel are
> valid/make any sense. I've attached outputs, they seem off the scale and
> it has been suggested to me that there is something wrong, although all
> sections of the analyses went swimmingly! Perhaps these values are valid
> and the constrained tree is wildly less likely than the other?
>
> I got per site log likelihoods from comparison of two trees generated
> using RAxML, one with and one without a topological constraint (tree
> file attached).
>
> I’ve looked through previous postings and thought that I may be
> experiencing the same problem as Gregory Mendez back in October 2015 –
> constrained group not strictly monophyletic (second tree in the file),
> but I’m not sure that it isn't (as Alexis says in the manual) that taxa
> not under a constraint are free to fall into any part of the tree.

You need to check the respective tree, in general the constraint trees
work correctly, but I have still not been able to fix the bug that
Gregory mentioned, which is very tricky.

If you use a comprehensive constraint tree (containing all taxa) I am
pretty sure my implementation is correct, if some taxa are missing in
the constraint tree then there might be a bug and I suggest you should
inspect your tree visually.

> I
> have rooted the tree on the appropriate outgroup and it makes no
> difference. The two TracemXXXX taxa are the ones that I "forced" into
> the same clade as all other 'T" taxa, but following ML optimisation they
> seem to be part of a trichotomy.
>
> For constraint RAxML was called as follows:
> raxmlHPC -f a -x 12479 -p 14507 -# 1000 -m GTRGAMMA -q geneparts.txt -s
> kalan87.phy -g constraint.tre -n a5
>
> I would really appreciate any advice on this.

Can you please send me all the input data to my personal email such that
I can check what is happening?

Alexis

>
> Many thanks
> Tracey
>
> --
> 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

[Alexandros Stamatakis]

Dear Tracey,

I just looked at the results of your comprehensive constraint tree
search and the ML tree you got, does indeed comply with the constraint
you specified.

Just keep in mind that those trees are unrooted to better understand the
issue.

In the constraint you have forced the two taxa Tracemxxxx to be
monophyletic with all other taxa starting by a T and this is indeed the
case, the interesting part of your tree looks like this:

-----some taxa starting with a C
|
----| |----The other T taxa
| |
----|
|------the two Tracemxxx taxa

So this does not seem to be the issue.

It might indeed happen that a tree search conducted under a topological
constraint may yield a likelihood that is far worse than the likelihood
of the unconstrained search.

This means that your constraint contradicts the signal from the data.

One reason could be that the sequences have been labelled by incorrect
sequence names.

Hope this helps,

Alexis