Thursday, January 10, 2013

Graphical illustration of some important concepts and terms

The following is about phylogenetic systematics and its detractors again. Please disregard if the topic does not interest you.

Some concepts and terms are often misunderstood, confused or conflated, and it seems useful to illustrate them in a concise way for later reference.

The above is a phylogeny, or a phylogenetic tree. It has a structure that is branching, i.e. tree-like, or in other words, it has a phylogenetic structure. Even if there were a few reticulation events, the structure would be mostly tree-like, and we would still call it a phylogeny.

Examples for systems that show a phylogenetic structure are groups of:
  • homologous sequence copies / alleles / haplotypes / chromosomes, with some reticulation possible due to a process that allows recombination between alleles (crossover)
  • chloroplasts and mitochondria
  • cells within one organism
  • individuals that belong to the same asexually reproducing species
  • biological species, with some reticulations possible due to allopolyploid speciation

The red area in the above phylogeny shows a monophyletic group. It covers an entire branch of a phylogeny.

The red and orange area here is a non-monophyletic group. It covers only part of a branch of a phylogeny.

With the same content of extant species, it can be either paraphyletic or polyphyletic depending on whether the orange part is included or not, which in turn depends only on the ancestral character state of the character used to define the group. That shows quite nicely that there is hardly any difference between paraphyletic and polyphyletic groups, a fact one should perhaps take into consideration when discussing the proposal that paraphyletic groups be given formal recognition, or even just when discussing the attempt by evolutionary systematists to redefine monophyletic to mean either monophyletic or paraphyletic. (This is often justified with the observation that paraphyletic groups also have a common ancestor, but well, the same is true of any two species on this planet, even Homo sapiens and Escherichia coli.)

In contrast, this is not a paraphyletic group. It is part of a polytomy, either a true one or, more likely, part of a phylogenetic tree that is not fully resolved. In reality, it could be a paraphyletic group, or it could be a monophyletic group, but at the moment we simply cannot tell, and so it is neither one nor the other but a we-don't-know-phyletic one. The technical term for this is metaphyletic.

Memo to evolutionary systematists: Please reconsider justifying the acceptance of paraphyletic taxa with polytomies, uncertainties or low bootstrap support; in those cases, the groups you want to circumscribe may even be monophyletic!
This illustration shows a network. It has a structure that is reticulating, also called a tokogenetic structure.

Examples for systems that show a tokogenetic structure are groups of:
  • individuals that belong to sexually reproducing species
  • (well, that is it as far as biological systematics is concerned, unless allopolyploid speciation were a good deal more frequent than can be made plausible)
The red area in the above network is not a paraphyletic group because there is no phylogenetic structure for it to be paraphyletic in. It may look paraphyletic if we force the data from this tokogenetic system into a tree-like representation, but in reality it isn't, and science is about describing and modelling reality.

Memo to evolutionary systematists: Please reconsider justifying paraphyletic taxa with reticulation. If reticulation is rare enough, monophyletic taxa work fine. If it is frequent enough for monophyletic taxa to become problematic, there aren't any paraphyletic ones either.

Memo to cladists who think that the sentence "this species is monophyletic" makes any sense whatsoever: This is why it doesn't. Unless you are talking about asexually reproducing species, of course, because groups of individuals from those do have phylogenetic structure.

So is there a name for a group like the one in the last illustration? I am not sure if somebody has already bothered to invent these terms, but paratoketic sounds about right as a counterpart to the terms ending in -phyletic. A group containing all descendants of the common ancestor would then obviously be monotoketic.

Finally, consider the big picture: You have gene copies forming a phylogenetic structure within individuals, within populations, within species. You have organelle genomes forming a phylogenetic structure within individuals, within populations, within species. You have cell lines forming a phylogenetic structure within an individual. You have individuals and populations of sexually reproducing species forming tokogenetic structures, and individuals and populations of asexually reproducing species forming phylogenetic structures. And finally, you have all the species on the planet forming a (mostly) phylogenetic structure. All intertwined, on top of or within each other in a picture that is breathtakingly complex. But phylogenetic systematics is only about the last of all of these: discovering and naming monophyletic groups of species. Not monophyletic groups of gene copies, not monophyletic groups of individuals, and so on.


  1. Alex, I appreciate your comments on the basics of cladistic analysis. Would you agree with this statement: "all polyphyletic groups are necessarily paraphyletic, but not all paraphyletic groups are polyphyletic"? Thanks.

  2. Not sure if that is a very productive way of looking at it. Every non-monophyletic group of extant species can be paraphyletic or polyphyletic merely depending merely on whether their most recent common ancestor is assigned to the group or not, and that in turn is merely a question of what that ancestor is reconstructed to have looked like...

  3. Well, usually we don't directly observe the common ancestors, we infer them. A polyphyletic grouping could be corrected to monophyletic in one of two ways: split the group until each remaining group is monophyletic, or include more branches of the tree until the most recent common ancestor and all of its descendants is part of the ingroup. In the latter method of eliminating polyphyly, it seems apparent that a polyphyletic group is not only missing the common ancestor, but also some descendants. Missing descendants is the definition of a paraphyletic group. I'm seeking a way to help undergraduates in introductory biology courses understand the distinction between monophyly, paraphyly and polyphyly. However, I am not a phylogeneticist myself, so I'm trying to improve my understanding of the concept at the same time.

  4. The point is that a paraphyletic group is also missing some of the descendants. Again, the only difference between a para and a poly group is in whether the ancestors that are also assigned to the 'group' form an unbroken line connecting all extant members or not, as in the figures above. Another way of addressing the issue is by asking if the character used to recognise the 'group' has arisen only once in the ancestor and then been transformed (or lost) in the nested members that are excluded, or if it has arisen several times in multiple ancestors of the 'group': in the first case, it is para, in the second it is poly. There are probably also other illustrations of this stuff on the internet that you could check out.

  5. I said in my last comment that missing descendants in the definition of a paraphyletic group, so I think I understand that point. If a polyphyletic grouping is not only missing ancestors but also unavoidably missing descendants as a result, then it would be both paraphyletic and polyphyletic. If you think it's the case that a polyphyletic group is also always paraphyletic, then the answer to the question I asked in my first comment would be "yes".

    And I've read the first chapter of The Compleat Cladist, which introduces these concepts, but was interested in whether an expert in the field agreed with my interpretation.

  6. Ah, I see what you are getting at. I guess it just isn't very customary to nest the definitions like this; the usual way of putting it would be:
    1. monophyletic
    2. non-monophyletic
    2a. parapyletic
    2b. polyphyletic
    (3. Could be either but we don't know yet: metaphyletic)

    Most pro-paraphyly people prefer the following:
    1. monophyletic
    1a. holophyletic (= monophyletic sensu Hennig)
    1b. paraphyletic
    2. polyphyletic

    ...although interestingly Richard Zander sometimes claims that cladists would not see a big difference between para and poly although there really is one.

  7. Alex, thanks for taking the time to reply to me on this. I'm going to try to find time to read at least the first chapter of The Compleat Cladist again, and possibly some other references as well to try to straighten things out in my mind. Possibly I'm misunderstanding or missing some aspects of the core concepts, but I think I lean toward your first set of nesting/ dichotomies of the concepts. I'm also not certain that paraphyly can be avoided entirely due to the time dimension and cladogenesis. I do think that polyphyly can be avoided and should be avoided.

  8. I don't really follow; paraphyly can be avoided by not recognising paraphyletic taxa. Not sure what you mean with 'due to time and cladogenesis'. The argument is usually that paraphyly cannot be avoided due to reticulation, but as I tried to explain in the above post any situation in which there is no monophyly is a situation where there is no paraphyly either...

  9. If I had a good argument, I'd make it. At the moment, it's just a half-formed impression. I'm trying to more fully understand the issues. I could be wrong. Knowing that species aren't expected to be monophyletic might obviate the issue, though. I'm still mulling it over.