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