Constraint Trees

[Will Pearse]

Hello,

I've been trying to specify a complete constraint, but I can't seem to get it to work. I'm very confused, as I've been able to do this in the past and only seem to be having trouble now!

I downloaded the latest version of RAxML (7.3.2, June 2012) and compiled with 'Makefile.SSE3.gcc'. I've taken the 'hands on' dataset for this example, and you can download everything I use here (https://dl.dropbox.com/u/4149392/constraints.zip). When I run:

./raxmlHPC-SSE3 -p12345 -r fullConstraint.txt -m GTRGAMMA -s dna.phy -n constrained

I find the contraint tree has been ignored - for example, I constrained the rat and the loach to be together, and yet they're not. The same happens if I use an incomplete, multifurcating constraint, e.g.:

./raxmlHPC-SSE3 -p12345 -g partialConstraint.txt  -m GTRGAMMA -s dna.phy -n partial

I'm sure I've done something incredibly stupid, and I'm sorry for that, but I'd appreciate knowing what that is!

Thanks,

Will Pearse

[Akifumi S. Tanabe]

Hello Will,

Use the following commands for optimizing parameters under a fixed
tree topology.

raxmlHPC -f e -t fullConstraint.txt -m GTRGAMMA -s dna.phy -n complete
or
raxmlHPC -f d -g fullConstraint.txt -m GTRGAMMA -s dna.phy -n complete2 -p 12345

Note that your fullConstraint.txt contains multifurcation on root
point.

Use the following constraint instead of partialConstraint.txt. This
may raise your expected tree.

((Frog,(Cow,Rat,Carp)),Chicken,Human,Loach,Mouse,Seal,Whale);

Note that the behavior of RAxML may not be a bug.

Best regards,

On Tue, 17 Jul 2012 07:42:38 -0700 (PDT)
Will Pearse <will....@gmail.com> wrote in
<357c0f25-11b1-4617...@googlegroups.com>

--
Akifumi S. Tanabe <akifumi...@gmail.com>, Ph.D.
Graduate School of Global Environmental Studies, Kyoto University

[Will Pearse]

Hello,

Fantastic, thank you!

Will

[Elisabet]

Hi

I am having the same problem with one of my datasets, the constraint seems to be ignored. I get an unresolved tree with one of the groups I defined as monophyletic being a polytomy shared with one more group. I have tried several different ways of defining the constraint, but it ends up the same. For other datasets this has been working fine, in the same version and all. Version 7.2.8

Heres how I've called RAxML
raxmlHPC-SSE3 -p 1234 -g Pha_macounii_mono.txt -m GTRGAMMA -s tef.phy -n Pha_macounii_mono3
 I also tried
raxmlHPC -f d -p 12345 -g Pha_macounii_mono.txt -m GTRGAMMA -s tef.phy -n Pha_macounii_mono4

I attached the constraint and the output tree. Any suggestions on how to improve this are welcome.

Best regards
Elisabet

[Tony Gamble]

Elisabet,
You might want to view your tree without branch lengths, i.e. as a cladogram (easily done in Figtree), to see if the constraint is enforced. My guess is that the branch leading to the two clades in green is there but it is extremely short making it appear as a polytomy.
Tony

----------
Tony Gamble, Ph.D.
Postdoctoral Researcher
Department of Genetics, Cell Biology and Development
University of Minnesota
6-160 Jackson Hall
321 Church St SE
Minneapolis MN 55455
USA
Phone: 612-626-4259
E-mail: gamb...@umn.edu
Web: www.tc.umn.edu/~gambl007

> <Pha_macounii_mono.txt><Pha_macounii_mono.pdf>

[Elisabet]

Tony Thanks!

You were absolutely right.

Elisabet

[Alexandros Stamatakis]

Dear Elisabet,

Please update to the latest RAxML version from github.

Also search for previous discussions on constraint trees in the google
group please.

All the best,

Alexis

--
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

[Ingo Michalak]

Dear Alexis,

sorry for continuing in this endless story but might it be, that topologic constraints are not implemented in the fast tree search (-f E)?

I can successfully constrain with a multifurcating tree in ML searches and BS analyses, but it fails in the fast tree search. 

Here is what I did (using your example alignment, and RAxML version 7.3.2 from July, 31st; on win7):

raxmlHPC.exe -f E -m GTRGAMMA -g constr.txt -n ctest -p 1234 -s dna.phy

with constr.test being:

((Mouse, Rat, Chicken), (Human, Loach, Seal), (Whale, Carp), Cow, Frog);

And I got:

((Loach,Carp),((Whale,(Human,(Chicken,(Rat,Mouse)))),(Frog,Seal)),Cow);

However, in the ML search it worked (with the same files) to run:

raxmlHPC.exe -f d -N 1 -m GTRGAMMA -g constr.txt -n test -p 1234 -s dna.phy

I got (as expected):

(((Chicken,(Rat,Mouse)),(((Human,Seal),Loach),Frog)),(Whale,Carp),Cow);

Did I miss something?


Best regards

Ingo

[Alexandros Stamatakis]

Hi Ingo,

You didn't miss something, in fact, the fast tree search option simply
ignores constraints.

I will either implement them for the fast search as well or add an error
message that will abort RAxML in case constraints are specified
together with the fast tree search, depending on my time.

Alexis