Help plot phylogenetic boostrap values from ClustalW

Nicolas Le Nov=E8re lenov at pasteur.fr
Mon Apr 20 13:52:41 EST 1998



Jan Hansen wrote:

> How do I get the boostrap confidence values for each node
> plotted on the graph, together with a bar indicating how
> much 10% sequence divergence is in horisontal direction?
> Neither drawgram nor drawtree support such feature?
>
> My plots should be in postscript or other format.
> Any  such phylo plot programs available for UNIX

As far as I know, if you bootstrap an alignment and make repeated
analyses,
you loose all distance information. Indeed, each reconstruction use a
different alignment. Except if your divergence rates are constant across
the
whole sequence length, you will get a different distance each time.
The final consensus tree gives you a topology: "this node was found in
xx% of the cases". The length disappeared. A possible solution is
to plot the corresponding boostrap values on the actual nodes. You loose
the
topological information coming from the multiple analysis, however you
will
get very low bootstrap scores on the wrong nodes.
Another possibility is the "mirrored trees" with the OTUs in the middle,

the actual NJ tree on the left and the consensus tree (with bootstrap
values) on the right. To perform  these two figures, the best is maybe
to
used a drawing software. I played a lot with Drawgram, Drawtree,
Treetool
and Treeview, but finally, I drawed all my trees with Canvas and
Illustrator.

--
Nicolas Le Nov=E8re, Neurobiologie Mol=E9culaire, INSTITUT PASTEUR
25, rue du DR ROUX, 75724 Paris cedex
Ph: 33-(0)1-45-68-88-44, Fax: 33-(0)1-45-68-88-44, email:
lenov at pasteur.fr
http://www.pasteur.fr/units/neubiomol/





More information about the Mol-evol mailing list