A reader commented by email about my criticism of the PLoS ONE article that inferred a multigene phylogeny of eukaryotes, with Chlamydomonas reinhardtii as the outgroup (“A cautionary tale on reading phylogenetic trees“).

Although you are of course correct to complain about nearly everything in the paper (esp. re “basal” and node rotations), and I am sure the tree is wrong in more ways than it is right, I think you might reconsider or put in context complaints about the “provides a link between”. My thought is simply that if one has a long branch between two nodes in a tree, if you add a taxon group that branches off in the middle of this long branch, then it does, in a sense, provide a “link” between these two nodes. A more proper way to put it is that it provides information concerning the ancestral state at the two original nodes (i.e., may substantially modify the posterior probability of the states at the two nodes). I doubt that the authors mean it in this sense, but in the general context of teaching people about trees, I would want students to understand this.

This is in response to my objection to the claim that the choanoflagellate S. rosetta “…provided a link between fungal and animal species.”:

No, it doesn’t. S. rosetta is sister to the animals. It it exactly as distantly related to the fungi as are all the animals:

Take a close look at that figure. The red circle represents the divergence between S. rosetta and the fungi. It also represents the divergence between humans and fungi, between cows and fungi, and between mosquitos and fungi. S. rosetta is exactly as closely related to fungi as animals are.

T. gondii provides a link between fungal and plant species…

Again, and for the same reasons, no. T. gondii diverged from plants at the same time the animals and fungi did, at the blue circle.

Well…I agree. Mostly.

My objection was because to me, ‘link between’ implies that S. rosetta is somehow intermediate between fungi and animals, which it absolutely is not; S. rosetta and animals are equally distant from (share the same MRCA with) fungi. Given node rotation, it would be totally equivalent to say ‘animals provide a link between S. rosetta and fungi’.

My correspondent is right, though, that Salpingoeca rosetta is useful in inferring ancestral states. My concern, as always, is that we distinguish between extant relatives and ancestors. Modern choanoflagellates are no more ancestors of animals than chimpanzees are ancestors of humans. But modern choanoflagellates, when understood in their proper phylogenetic context, give important clues about the nature of animal ancestors. Traits that are shared between choanoflagellates and animals, for example, were probably present in the most recent common ancestor of animals and choanoflagellates. Traits shared by animals, choanoflagellates, and fungi were probably present in the most recent common ancestor of those groups (the red circle in the figure above).

I tend to prefer strictly cladistic descriptions of phylogenetic trees, e.g. animals are sister to S. rosetta, fungi are sister to S. rosetta + animals. But the authors aren’t bound to my preferences; I’ll be the first to admit that I can get pretty pedantic about this stuff. If we interpret ‘provide[s] a link between fungal and animal species’ as ‘gives insight into the ancestors of animals and fungi’, then it’s true. As my correspondent points out, it’s unclear exactly what the authors meant by this, but in principle ‘link between’ can be a useful shorthand.