Promoting Plants at the Expense of Fungi?

Ryan Gregory, over at Genomicron, found an interesting tidbit on the Discovery Channel website entitled Plants and Animals: Long-Lost Relatives?. He is surprised to read that

"organisms such as fungi should be given a demotion — placed further from animals on the tree — while green plants should get a leg up."

This is the case, apparently according to a recently-published paper from John Stiller at East Carolina University.

Gregory is (rightfully) annoyed at the suggestion of any group "getting "demoted" one way or another because this idea of rank (was) should have been abandoned 150 years ago." and that was the main point of his post. Ryan has been (again, rightfully) critical of science reporting and this doozey has all of the right parts, as detailed in his Anatomy of a bad science story.

However, there have been a series of comments about the veracity of the Stiller paper, so I thought I would make a few of my own here. First of all, it should be duly noted that the paper in question, Plastid endosymbiosis, genome evolution and the origin of green plants, is explicitly labelled as "Opinion"; this is on top of the fact that it is published in the review journal TRENDS in Plant Science—not a primary research venue. Neither of these facts are damning to the work, but they certainly suggest caution in reporting the findings.

Second, it is fair to say that this hypothesis (wrongly called "Stiller's theory" by the Discovery story) is way out of the mainstream of current thought. Again, this alone should neither preclude the publication nor, by itself, lend it to immediate scorn. It's great to see such examples of how science works in the marketplace of ideas. But we all know that just because something gets published, does not mean it's right.

In any case, there are two main components that Stiller argues in this paper:

1. that the "Plantae" [Viridiplantae (green algae & land plants), Rhodphyta (red algae) and Glaucocystophyta] are not a monophyletic group, and
2. that the Viridiplantae are more closely related to animals than are fungi.

The data on "Plantae" monophyly are certainly not so compelling as to rule out other possible answers. But recent work (here and here) is increasingly providing support for this relationship. Stiller has been a long-standing contributor to the literature on non-monophyly of green+red algae. I have worked on some of the same molecules that he has (i.e., RNA pol II) and he might have a point for there not being strong support in favor of "Plantae". However, I don't think that there is an alternative that garners anywhere near the consistent support that "Plantae" gets in multigene trees.

The data on the sisterhood of animals and fungi grouping to the exclusion of plants (e.g., Viridiplantae) is about as solid as deep relationships among eukaryotes can be. A nice summary of these results was recently presented in PLoS Genetics. Since the paper is Open Access, I won't repeat the findings here (full disclosure: although I am not an author of this paper, I do collaborate with them). Suffice it to say that there is not even a hint of evidence that Viridiplantae is closer to animals that are fungi as suggested by Stiller. Note also that these authors are not particularly bullish on "Plantae" either.

Stiller cites the presence of certain enzymes and protein domains in both plants and animals (apparently absent in the fungi) as evidence for a closer relationship of plants to animals. But the problem is that such things only have to be lost once in the fungal lineage (if they have been lost at all) to make these cases complete non-sequiturs. To explain why plants don't fit in with animals, Stiller would rather invoke some sort of bias in the data. Although such biases may exist, I am very cautious about invoking an entire reworking of the tree based on them.

Stiller ends with four "Future Perspectives" of which I find the following most telling:

"There should be no a priori assumption that the strongest tree-building signal in a given data set reflects evolutionary history rather than bias in the data."

I'll file that in the category of "Things that make you say Hmmm".