Estimating clock-like branch lengths

Korbinian Strimmer strimmer at zi.biologie.uni-muenchen.de
Fri Jan 12 07:04:43 EST 1996


Given a possibly multifurcating tree T, a model of sequence evolution M,
and the data set D (nucleotides or amino acids), how can I assign branch
lengths to T in such a way that T obeys a molecular clock?

The preferred way is, of course, maximum likelihood.  If T is strictly
bifurcating, if the data are nucleotides, and if the model of sequence
evolution is the F84 model, I use maximum likelihood as implemented in
DNAMLK (PHYLIP).

Which programs are available if T is multifurcating?  Or if D are amino acids?

(NucML or ProtML provide many models and can be applied to multifurcating trees,
but they don't return a clock-like tree ... !)


Korbinian Strimmer



More information about the Mol-evol mailing list