Several enhancements have been added to the all-against-all
e-mail server at ETHZ: besides computing phylogenetic trees
and multiple alignments, the server now performs two
additional functions:
1) Computation of the split decomposition of the estimated
evolutionary distances according to [1]. Split decomposition
is a new way of visualizing estimated evolutionary distances
which do not fit perfectly on a phylogenetic tree. The key
idea is not to separate a single taxon from all others, but
to separate a set of taxa from its complement.
The server draws (in PostScript) the split decomposition as a
graph similar to a tree. The simplest example with just four
taxa A,B,C,D will look like
A B
\ /
\__________________/
| |
| |3
|__________________|
/ 14 \
/ \
C D
This graph directly visualizes two possible trees and the fact
. .
/ \ / \
that the tree /\ /\ is more likely than /\ /\:
AC BD AB CD
it easier to cut the graph into (AC)(BD) with a cut of length 3
than cutting it into (AB)(CD) with a cut of length 14.
2) Computation of variation indices over multiple alignments.
The classical Kabat-Wu index [2] as well as two new indices
based on mutation probabilities are available:
For ProbIndex, the probability of the mutations implied by the
phylogenetic tree is computed for each position of the multiple
alignment. The negative logarithm to base 10 of this probability
is printed and plotted as a histogram in PostScript.
For ScaleIndex, a scale factor is computed for each position of
the multiple alignment such that the probability of the mutations
implied by the phylogenetic tree with pam distances scaled by this
factor is maximized. The logarithm to base 10 of this scale factor
is printed and plotted as a histogram in PostScript.
For more information, send the message "Help" or "Help AllAll" to
cbrg at inf.ethz.ch
[1] H.J. Bandelt, A.W.M. Dress: Split decomposition: a new and
useful approach to phylogenetic analysis of distance data.
Molecular phylogenetics and evolution, to appear.
[2] T.T. Wu, E.A. Kabat: An analysis of the sequences of the
variable regions of Bence Jones proteins and myeloma light
chains and their implications for antibody complementarity. J.
Exp. Med. 132(1970):211-250.
