Keith Robison asks about a recent paper by Horai et al. in J. Mol. Evol.
> Did they do their analysis correctly? Are the numbers really as good as
> they claim? Can this be considered bedrock proof for a chimp-human clade?
Since I haven't seen the paper in question yet I can't say. I do wnat to
remind everyone that even if the answer to the first two questions is a
resounding "Yes" that doesn't mean that the answer to the last one is also
"Yes". Answering the first two questions yes would allow us to conclude that
there is very strong support for a chimp mitochondrion-human mitochondrion
clade, but that isn't necessarily the same thing as a chimp-human clade.
Gene trees do not have to match species trees.
If, however, the analysis includes multiple human sequences, multiple
chimp sequences, multiple gorilla sequences, etc. and if each species forms
a clade in the analysis, then we have reasonable evidence that the gene
tree will be congruent with the species tree. If we have only one sequence
per species, we must be more cautious in our conclusions.