More on the 'cladistics' of sequences
skala at INCOMA.CZ
Wed Jun 9 08:52:58 CDT 2004
Pierre about multiple outgroups:
.. the congruence criterion optimizing the topology
and rooting for all characters will have done the trick.
The internal topological arrangement of outgroups don't seem to be a
problem for me: the overall optimal one is taken into account in the
1. Even in the case we have one outgroup terminal (and hence no character-state variation can be observed there) the character states found in outgroup (and believed generally to be plesiomorphic relative in the group studied) *can* be homoplastic (why not when we are 'observing' homoplasies in the group itself). There is however hardly a way to include this in the analysis (polarization).
2. This is related to the internal topology of the (potential) multiple outgroups. "Congruence" among the outgroups you are referring to is simply a similarity-based operation ("phenetic" one) since character state polarization is (in the typical case) unknown. Imagine that the outgroup terminals (or subset of them) make a monophyletic group sister to the group under the study. Then, there can be homoplasies basal to these outgroup taxa or to some of them. Consequently, either no character-state variation is observed in the outgroup (when homoplasies are placed basally) or homoplasies prevail and will be traced as those "most informative" (when arising in a sub-clade of the outgroup terminals). That is why internal topology of the outgroups *is* a problem.
More information about the Taxacom