Friday, November 23, 2012

Introduction to Phylogenetic Systematics

As indicated in my first post, one of my professional interests is Phylogenetic Systematics (subsequently PS). As a reference for future discussion, this and perhaps a few follow-ups will provide a concise introduction. However, there is a limit to how concise one can be and still make sense, so this post will necessarily be quite long.

First, what is PS? 

PS is a school of biological classification that was developed by the German entomologist Willi Hennig (1950, 1966). After a long controversy, it has become the dominant paradigm in systematic biology since about the 1980ies. Some opposition still exists - perhaps more so in botany than in zoology - but that is something I will cover another time.

Instead, the details. Hennig's first relevant insight was that classifications developed by biologists are not actually based on the examination of entire species, populations or individuals, but on the examination of snapshots of (some aspect of) organisms at a particular time; he called these "semaphoronts". Examples of semaphoronts are stuffed animals, needle-mounted insects, fossils, photographs, blood samples or root tips, or indeed many other types of biological samples.

The second, and more directly relevant, insight was that these semaphoronts can be related to each other in three different ways:

1. They can be from different stages in the development of one individual. In this case the relationship of the semaphoronts is called ontogenetic, and the mental model for that relationship is a strict linear arrangement, from earlier to later. Examples would be photographs of you as a baby and of you as an adult, or herbarium specimens taken from the same tree in 1920 and 1995.

2. If they are from different individuals of the same sexually reproducing biological species, the sempharonts have a (predominantly) tokogenetic relationship. In this case, the model of the relationship is a network with constant splits and reticulations, as every individual has two direct ancestors and may have multiple descendants. Examples of semaphoronts with a tokogenetic relationship would be a blood sample from you and one from me, or two stuffed specimens of brown bears.

3. Finally, if they are from different species, the semaphoronts have a (predominantly) phylogenetic relationship. It is modelled as tree-like, with constant splits but rare or no reticulations: in the course of evolution, species are assumed to potentially have several descendant species but usually only one immediately ancestral species*. Examples are the relationship between a herbarium specimen of the dogrose and one of the common thistle, or between a photograph of you and one of an oak tree.

From these considerations, Hennig deduced that different approaches of classification have to be adopted for the different relationships that semaphoronts can have with another. His interest, as the "phylogenetic" in PS already indicates, was mainly with the classification above the species level, where semaphoronts have phylogenetic relationships. So from now on we assume that we have grouped our semaphoronts into species, and concern ourselves with classifying several species on a phylogenetic tree into more or less inclusive groups, or supra-specific taxa. Assume also for the moment that we know the phylogenetic tree - because how to infer that is definitely beyond the scope of this post.

The next bit of terminology that we have to consider concerns how to describe the different types of phylogenetic relationships that species can have. Hennig distinguished three types of groups one can circumscribe on a phylogeny:

1. Groups of species that include all descendants of one common ancestral species are called monophyletic. They represent an entire branch of the tree of life that could be snipped off with just one cut. Monophyletic groups can be recognized by sharing a character that all the member species originally inherited from the common ancestor (although it may sometimes subsequently be lost again), and that character is then called a synapomorphy. For example, the possession of four legs is a synapomorphy for the land-living vertebrates, AKA tetrapods, and the secondary loss of that character is a synapomorphy for the snakes.

2. Groups of species that include some but not all descendants of one common ancestral species are called paraphyletic. You can visualize them as one branch of the tree that had one or more of its sub-branches removed, i.e. you need at least two secateur cuts to get them from the tree of life. Paraphyletic groups are defined based on old character states (symplesiomorphies) that have been lost in a phylogenetically nested group that would have to be included in the paraphyletic group to make it monophyletic. For example, the invertebrates are a paraphyletic group because they are only defined by not having evolved a spine, which is a synapomorphy for the (monophyletic) vertebrates.

3. Groups comprised of species that are not directly related are called polyphyletic. In this case you would snip off several distant branches from the tree and throw them onto an arbitrary heap. Polyphyletic groups are defined based on characters that have evolved in parallel, but they are usually more easily recognized as artificial. An obviously paraphyletic group would, for example, be that of all winged animals, including insects, pterodactyls, birds and bats.

Illustration of mono-, para- and polyphyletic groups on a cladogram, from Wikimedia Commons; original work by Petter Bøckman, revised by Peter Brown. Personally I don't think much of the Wikipedia articles on the topic, however.

And here is now finally the crux of phylogenetic systematics: Its criterion for classification is to recognize only monophyletic groups and to reject all non-monophyletic groups. Why? There are several unrelated reasons that make this criterion so important:
  • A classification containing only monophyletic groups is the most natural one. Non-monophyletic groups always contain species that are more closely related to species outside of the group than they are to any other member of their group. A well-known example are the dinosaurs and the birds, which were traditionally treated as separate groups on the same level but now as the former containing the latter. If you classify the birds as apart from the dinosaurs out of which they evolved, there will always be a dinosaur that is more closely related to the birds than it is to any other dinosaur**.
  • Monophyly is an objective and testable criterion. Indeed it is the only one that has ever been suggested and thus the only alternative to deciding on the validity of biological classifications with popularity contests or shouting matches. In other words, only phylogenetic systematics makes classification and taxonomy a science.
  • Monophyly is also a universal criterion that works for all groups of organisms as long as they show phylogenetic structure. Even if those still in opposition to phylogenetic systematics could ever come up with a definition of "different enough" to classify paraphyletic groups of plants apart from other plants nested within them, the same definition would not work for, say, insects, which of course have a very different morphology.
  • From a practical perspective, monophyletic groups are also much more useful than non-monophyletic ones. Because they include all descendants of one ancestor, it is reasonable to assume that they could have more synapomorphies than the ones that we have already discovered. So when you know one plant that produces a commercially interesting biochemical compound, it makes sense to search for other plants with more of the same or with similar biochemistry among its relatives. If it is classified into a non-monophyletic group, and you rely on that classification, you will overlook at least some of its interesting relatives, and if you are unlucky perhaps even all of its closest relatives.
If I made any conceptual mistakes, please do not hesitate to point them out. The next post will deal with a common misconception about Phylogenetic Systematics.


Hennig W, 1950. Grundzüge einer Theorie der Phylogenetischen Systematik. Berlin: Deutscher Zentralverlag.
Hennig W, 1966. Phylogenetic Systematics. Urbana: University of Illinois Press.


*)  Here, things are a bit more complicated: Especially in plants, species are sometimes allopolyploid and thus the result of a reticulation event between two ancestral species. Conversely, semaphoronts from an entirely asexually reproducing species also have a purely phylogenetic relationship. But that is beyond the scope of the present post.

**) Essentially, a paraphyletic group is the biological equivalent of grouping humans by their religious views into Catholics and non-Catholics; obviously Anglicans, for example, are much more similar in their beliefs to Catholics than they are to many other non-Catholics, like Hindus or Buddhists, and thus that classification would be artificial.

No comments:

Post a Comment