In response to a message posted on methods and reagents, but more appropriate
for the molecular evolution BB:
>Hey,
>I`m working with the programm package HUSAR in Heidelberg, especially
>with the subprogramm TREE.(D.F. Feng and R.F. Doolittle, J.Mol.Evol.
>25,351(1987)).
>I've made a phylogenetic tree of 9 aa-sequences and it had been as I've
>expected it.
>The branches of the tree are of different length with numbers, characterising
>the branch length.
>Now I need a basis for explaining the different numbers. All I know is, that
>they represent distance scores, but I do'nt know their exact meaning.
>Does anyone know if these numbers are relative units, or are they ment as
>mutation events?
>Thank's
>Stefan
You might refer to a newer set of articles on this TREE program:
They are both in Methods in Enzymology, volume 183. One by Feng and Doolittle,
(p 375-387) defines the distance you are interested in as:
D= -ln ((Sreal-Srandom)/(Sident-Srandom)) X 100.
The Similarity (S) scores are based on the Dayhoff matrix.
You might also consider another distance measure,
based on PAPA (parsimony after progressive alignment), which is described by
Doolitle and Feng in the same volume (p 659-669). You should test to see if the
tree from the PAPA distances has the same branching order as the matrix based t
ree.
Give these papers a careful reading and hopefully you will better understand
the program that you are using.
One explanation of the different distance values is that the nine
sequences have evolved at variable rates. Phylogenetic trees with equal
branch lengths are a very rare sight.
Mike
*********************** Michael W. Smith **************************
* Salk Institute for Biological Studies *
* Molecular Genetics Laboratory phone: (619)453-4100, ext 376 *
* P. O. Box 85800 fax: (619)558-9513 *
* San Diego, CA 92138 E-mail: mws at molly.sdsc.edu *
*******************************************************************
