Phylogeny Friday - 15 September 2006

A few weeks ago I introduced the tree of life, albeit to some criticisms. The following week I zoomed in on one branch of that tree, the eukaryotes. I pointed out that animals were a mere twig in the eukaryotic tree, yet they have been the focus of a large amount of biological research. This disproportionate attention is due in part to our ignorance regarding the majority of eukaryotic taxa. We have only become aware of many of the eukaryotes recently, so we have a lot of catching up to do in order to understand their evolutionary relationships.

Because animals are the best studied eukaryotes, we have a decent idea of the different animal phyla. However, there is still much controversy regarding the relationships of these phyla. Before we get any further, allow me to present one version of the animal phylogeny (based on morphological characters):

The two outgroups are plants and fungi, and their positions are contradicted by those in the eukaryotic phylogeny shown previously. The purpose of this tree, however, is not to show you the current hypothesis of the relationship of animals, but to introduce some major groups. In this tree, the first split within the animals occurs between Poriferans (sponges) and everything else. Next, the radially symmetrical animals (Cnidarians and Ctenophores, which include jellies, anemones, corals, hydra, and comb jellies) split from the bilaterally symmetrical ones. The Bilaterans can be split up based on types of body cavities. Bilaterans lacking a true body cavity (Acoelomates) -- a group containing the flat worms -- branch off first. Next, the Pseudocoelomates (roundworms and others) split from the Coelomates. Finally, the Coelomates are split into Protostomes (mollusks, segmented worms, arthropods, and others), Deuterostomes (vertebrates, sea stars, and others), and Lophophorates (made up of a bunch of animals you've never heard of). Don't worry if you're not familiar with any of the terminology from the previous paragraph; it's either not important or will be defined later.

One Molecule, Many Changes

Molecular data have changed not only our understanding of the relationship between animals, plants and fungi, but also the branching order and structure of the animal tree. A paradigm shift occurred when ribosomal RNA (rRNA) genes were sequenced in many different animals. The phylogeny created using one rRNA gene gave a very different tree than that based on morphology:

As with the morphological tree, the Poriferans (sponges), Cnidarians (jellies, anemones, etc.) and Ctenophores (comb jellies) branch off first. The relationship of the Bilaterans, however, is quite different. Rather than being split based on body cavity (ie, coelom), the Bilaterans are first split into Protostomes and Deuterostomes (this refers to how the mouth and anus are formed during early embryogenesis). The Protostomes are further split into two new taxa: the Lophotrochozoans (containing the Lophophorates and other taxa) and the Ecdysozoans (molting animals). The branching orders within these two taxa are unresolved. The molecular data resulted in quite a reshuffling of the animal tree, which can be seen by comparing the color coding in the two previous trees. These trees were taken from this review.

Whole Genomes Swap Branches

Why use a single gene when you can use a whole genome? In the early twenty-first century (ie, a few years ago), the first three animal genomes were sequenced: human (a Deuterostome), Drosophila melanogaster (an Arthropod), and Caenorhabditis elegans (a Nematode or roundworm). The evolutionary relationship of these three species could be hypothesized using data from these genomes, which would shed light on the branching order of the animal phyla. Using over 100 genes, the following tree garnered the greatest support [link]:

This tree agrees with the previous tree based on morphology, contradicting the rRNA tree. Of course, it's obviously better because it's based on many genes, rather than a single gene. Or is it? Keep in mind that the rRNA tree sampled from many more taxa, which is important when doing phylogenetic reconstruction. This especially holds true over long evolutionary distances. But the rRNA tree is based on a single gene, and it is very bad practice to develop a species tree using a single gene tree. It turns out, the biggest problem facing the reconstruction of the animal phylogeny are the short internal branch lengths in the tree and the long external branches going out to the individual phyla. This causes a problem for both approaches that use a single gene (due to lineage sorting) and those that don't sample from enough taxa (due to long branch attraction).

A New Solution

TThere is an obvious solution to the aforementioned problems: sequence more genes from more animal taxa. Of course, this costs money and requires an intelligent plan of attack to best use the required funding. A paper in the pipeline at Trends in Ecology & Evolution presents an approach toward resolving the animal phylogeny. Here is how they represent our current understanding of animal evolution:

Figure I. A conservative consensus of the current state of the animal phylogeny. Phyla for which a large amount of data (genome or EST projects) is available are indicated in red. The dashed lines indicate controversial relationships.

The general structure of this tree looks a lot like the rRNA tree. In fact, it's pretty well accepted that the Ecdysozoa and Lophotrochozoa are real taxa (see here for some multigene support for this tree). The authors suggest a sequencing strategy (I won't get into the details here) that will take advantage of multiple approaches to resolve different sections of the tree (this is already underway at the Tree of Life Project). Of special concern are the relationships within Ecdysozoa and Lophotrochozoa, which are pretty much unresolved.

And just in case you think I'm suggesting that molecular data trump morphological traits . . . well, they do and they don't. The morphological tree that now appears to be wrong (ie, the first tree shown) was incorrect because of the characters chosen for the phylogenetic reconstruction. But DNA data have problems too. Sampling a small subset of all taxa or very few genes will result in misleading phylogenies. Upon reexamination, the new tree (with three Bilateran subgroups: Deuterostomes, Lophotrochozoans, and Ecdysozoans) does jive with what we know about early development and morphology.

Adoutte A, Balavoine G, Lartillot N, Lespinet O, Prud'homme B, and de Rosa R. 2000. The new animal phylogeny: reliability and implications. Proc Natl Acad Sci USA. 97: 4453-4456. [LINK]

Blair JE, Ikeo K, Gojobori T, and Hedges SB. 2002. The evolutionary position of nematodes. BMC Evol Biol. 2: 7. doi: 10.1186/1471-2148-2-7

Philippe H and Telford MJ. 2006. Large-scale sequencing and the new animal phylogeny. Trends Ecol Evol. In press. doi: 10.1016/j.tree.2006.08.004