question on phylogenetic analysis

Alexandra Hiller alexandrahiller at
Wed Aug 18 08:26:04 EST 1999


I am trying to infer phylogenetic relationships using DNA
sequences. I am using a phylogenetic program  that computes
genetic distances from the sequences and one can choose two
options: to compute  distances using transitions as well as
transversions, that is, giving both types of mutations the same
weight, or to compute them by using transversions only. In my
results I observed that transitions outnumbered tranvsersions, so
I would like to give more weight to transversions than to
transitions, say 3:1 (as kind of standarized in most studies) but
I don't have this option in the program. I constructed two
phylogenetic trees with the neighbor-joining method, which uses
the comouted genetic distances to produce a tree. The first tree
was constructed using distances computed with transversions and
transitions and the second tree was constructed using distances
computed with transversions only. In the first tree two clusters
are formed with bootstrap values of aprox. 60-70%. The rest of
the species are left unclustered. In the second tree all species
are clustered: the same clusters as in the first tree are formed
with bootstrap values a little bit higher (70-75%), plus the rest
of te species are  clustered in other two groups. 
My question is: is the second tree, where only transversions are
considered, valid at all? Would the first tree, where transitions
and transversions are considered, more valid than the second one,
although no more weight is given to transversions?

Thank you!

Alexandra Hiller
Justus-Liebig University
Giessen, Germany
email: alexandrahiller at

More information about the Methods mailing list