Dear Evoldir Members,
Thank you very much for the answers!
Here I summarize them:
Alexandros Stamatakis:
you can do this with RAxML, given that you have an alignment
for the prot seqs.
Run RAxML (please get the latest version from GITHUB at:
https://github.com/stamatak/standard-RAxML) as follows:
./raxmlHPC-SSE3 -P proteinMatrixName -t referenceTree -s alignment -f x
-m PROTGAMMAWAG -n T1
Via -P you read in the protein substitution matrix:
-P Specify the file name of a user-defined AA (Protein)
substitution model. This file must contain
420 entries, the first 400 being the AA substitution rates
(this must be a symmetric matrix) and the
last 20 are the empirical base frequencies
-f x: compute pair-wise ML distances, ML model parameters will be
estimated on an MP
starting tree or a user-defined tree passed via "-t", only
allowed for GAMMA-based
models of rate heterogeneity
Don't be irritated by the WAG -n the model string -m PROTGAMMAWAG RAxML
will ignore WAG and
only use the information that the GAMMA model of rate heterogeneity is
to be used.
The -f x option will then just compute pair-wise ML distances between
all pairs of sequences.
The starting tree passed via -t is used to estimate the ML model
parameters (in the above case just the alpha parameter) that are then
used for calculating pair-wise distances.
*****
Gangolf Jobb:
Treefinder can. Use the "MAP" model, see manual page 28 (www.treefinder.de).
*****
Gwennaël Bataille:
I think the R software should allow this, since it is very flexible.
Books like "analysus of phylogenetics and evolution with R" by Emmanuel
Paradis could be useful for this, I suppose. Or alternatively, the
r-sig-phylo mailing list, where you can ask specific questions.
*****
Joe Felsenstein:
I think it is possible in PAML too. I have some Mathematica files that
calculate various of the required matrices.
We hope to add WAG and MtREV too to Protdist this year.
*****
Cameron Weadick:
You can use the aaratefile option in the PAML software package's CODEML
program to specify your own amino acid substitution rate matrix.
*****
Carolin Kosiol:
If you like to use PHYLIP, the AA substitution matrices need to be
encoded in the "protdist.c" file in the form of an eigen-decomposition
that is produced with Mathematica.
A description of eigen-decomposition and other methods to implement
substitution matrices can be found in the following article:
Kosiol C and Goldman N, 2005.
Different versions of the Dayhoff rate matrix.
Mol Bio Evol 22: 193-199.
I've tried Treefinder but unfortunately I was not able to import my
aminoacid distance matrix to this software and I wasn't able to detect
the reason why it didn't like the file format of the matrix.
Finally I've solved the problem with RAxML.
Bests,
Eszter Ari
2012-07-06 11:29 keltezéssel, evoldir@evol.biology.mcmaster.ca írta:
Dear Evoldir Members,
I would like to calculate distances (distance matrix) between protein
sequences using my own predefined aminoacid substitution matrix. Do you
know a software that allows this? The phylip protdist may allow it but I
was not able to find a description about the format of the substitution
matrix.
Thanks for the suggestions!
Eszter Ari
--
Ari, Eszter
assistant lecturer
Eotvos L. Univ., Dep. of Genetics
arieszter@gmail.com
+3613722500/8691
Ari Eszter