RAxML and "Six Dayhoff Groups"

[Hideyuki Miyazawa]

Dear all,

I am trying to recode the protein sequences into six Dayhoff groups (Hrdy et al. 2004)
and construct the phylogenetic tree.

Is it possible to do that with treating the six Dayhoff groups on the alignment
as the multi-state characters?

I know that it is able to do that with other phylogenetic analysis programs,
such as TREEFINDER, PhyloBayes, but, if possible, I want to do that with RAxML.

Please let me know.

Thanks,

Hideyuki, MIYAZAWA

[Alexandros Stamatakis]

Dear Hideyuki,

> I am trying to recode the protein sequences into six Dayhoff groups (Hrdy
> et al. 2004)
> and construct the phylogenetic tree.

Could you please send a link to this paper, I am not able to find it.

> Is it possible to do that with treating the six Dayhoff groups on the
> alignment
> as the multi-state characters?

I guess so, you'd have to map the groups to six multi-state characters
0,1,...,5 and then you can infer a tree using, for instance a GTR model
for multi-state characters, but I am not sure if this is what you want
to do because I am unaware of the model.

Alexis

>
> I know that it is able to do that with other phylogenetic analysis
> programs,
> such as TREEFINDER, PhyloBayes, but, if possible, I want to do that with
> RAxML.
>
> Please let me know.
>
> Thanks,
>
> Hideyuki, MIYAZAWA
>

--
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

[Akifumi S. Tanabe]

Hi Alexis and Hideyuki,

Hrdy et al. (2004) is available at the following URL.
http://dx.doi.org/10.1038/nature03149

Because GTR model assumes compositional equilibrium among OTUs, GTR
or similar models are not suitable for among-OTU compositionally highly
heterogeneous data set. "Data-recoding" such as RY-coding, AGY-coding,
or Dayhoff-coding is a technique for homogenizing character composition.

6-state Dayhoff coding should tlanslate

A, R, N, D, C, Q, E, G, H, I, L, K, M, F, P, S, T, W, Y, V

into

0, 1, 2, 2, 3, 2, 2, 0, 1, 4, 4, 1, 4, 5, 0, 0, 0, 5, 5, 4, respectively.

This translation might cancel out compositional heterogeneity, and we
can reconstruct ML tree applying MULTIGAMMA model by RAxML.

"pgrecodeseq" command in Phylogears can apply above translation. This
is available at http://www.fifthdimension.jp/products/phylogears/ . For
example, you can apply 6-state Dayhoff coding to an amino-acid sequence
data set by the following command.

pgrecodeseq --type=ANY "ARNDCQEGHILKMFPSTWYVX-01223220144145000554?" inputfile outputfile

This command can read FASTA, NEXUS, PHYLIP, relaxed-PHYLIP, and
Treefinder format sequence files.

Best regards,

On Fri, 11 Jul 2014 15:56:01 +0200
Alexandros Stamatakis <alexandros...@gmail.com> wrote in
<53BFECF...@gmail.com>

> > I am trying to recode the protein sequences into six Dayhoff groups (Hrdy
> > et al. 2004)
> > and construct the phylogenetic tree.
>
> Could you please send a link to this paper, I am not able to find it.
>
> > Is it possible to do that with treating the six Dayhoff groups on the
> > alignment
> > as the multi-state characters?
>
> I guess so, you'd have to map the groups to six multi-state characters 0,1,...,5 and then you can infer a tree using, for instance a GTR model for multi-state characters, but I am not sure if this is what you want to do because I am unaware of the model.

--
Akifumi S. Tanabe, Ph.D.
National Research Institute of Fisheries Science, Fisheries Research Agency,
2-12-4 Fuku-ura, Kanazawa-ku, Yokohama, Kanagawa 236-8648, Japan.

[Alexandros Stamatakis]

Dear Akifumi,

Thank you :-)

Alexis

[Hideyuki Miyazawa]

Dear Alexis and Akifumi,

Now, I am suspecting the compositional bias of amino acid sequences
as the cause of the unlikely tree topologies. And, I want to elucidate
how the Dayhoff recoding alters them.

Thank you very much for your rapidly answers, and helpful information on
Phylogears and the substitution model !!

Hideyuki MIYAZAWA


2014-07-12 0:50 GMT+09:00 Alexandros Stamatakis
<alexandros...@gmail.com>:

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