Wednesday, November 9, 2011

Some thoughts on the phylogenetic relationships between great apes and humans (and why morphology fails to answer this question)

Taxonomy isn’t a really satisfying branch of evolutionary biology, especially if you’re interested in the phylogenetic relationships between great apes, hominins and humans. You always tend to step on someone else’s toes and sooner or later this other person starts to harass you with critique. One of the standard arguments which are brought up against almost any kind of phylogenetic study is that the characters you used for your phylogenetic reconstruction are misleading, because they’re homoplasies. “Homoplasies” are characters which evolved independently in two closely related species and are usually your biggest enemy when it comes to taxonomic questions. 

fig. 1 Today's topic: The "correct" phylogenetic relationships between great apes and humans and why morphologic characters fail to reproduce these results.

This problem seems to be overabundant in Paleoanthropology, since I have yet to see a single taxonomic paper which isn’t immediately criticized for using homoplasious characters for their hypothesises. This is probably due to the reason that morphological characters in general are having a hard time reconstructing the “correct” phylogenetic relationships between great apes and humans (fig.1), which was established with molecular characters (e.g. Wood & Collard, 2002; Wood & Harrison, 2010; Strait & Grine, 2004). And if your characters aren’t able to reproduce the correct phylogeny between extant species, how can you expect them to get it right with fossil species?
Recently one of my Professors pointed me towards an interesting problem which made me wonder if those difficulties are really due to too many homplasious characters in your datasets.

Imagine a Population “P”
This Population has some characters which are polymorphic, which means that they exist in different states within this population. For the sake of simplicity I will only use two characters. The first character has either the state “+” or “-“
The second character has either the state “circle” or “square” (I’m too inept to put those symbols within this text –sorry).

Now this Population starts to split up.

Shortly after P1 split up from P, the rest of the original population splits into two additional subpopulations (P2 & P3).

Eventually all those three subpopulations will evolve into three different species

If we look at the pattern in which those Populations split from each other, we can see that P2 and P3 are closer related, because they got separated after P1 split away.
Now imagine that we weren’t able to witness this pattern and instead we have to reconstruct it with a cladistic analysis, which is usually the case if you want to know how extant species are related to each other. The problem here is that there is no clear signal. Instead your results depend highly on the character you’re using. Here we are in a situation where we have three equivocal hypothesises just because you had an polymorphic, ancestral population which split up so fast, that it wasn’t possible to fixate one character state within the whole population. Instead the ancestral polymorphic state was carried into those new subpopulations and the trouble we encounter if we want to use those characters for phylogenetic reconstruction is simply because of this incomplete lineage sorting and not because the character were homoplasious.

Of course I came up with this example, because we can see the exact same phenomenon if we look at molecular studies which dealt with the phylogenetic relationships between great apes and humans. Satta et al. (2000) found that out of 46 genetic markers, only 60% supported the “true” phylogeny (see fig. 1), the other 40% either supported a Human-Gorilla dichotomy, a Gorilla-Chimpanzee dichotomy, or a trichotomy.
Salen et al. (2003) came to similar result as they used “Alu-SINEs” to reconstruct the great ape phylogeny. Although Alu-SINEs are probably the best characters you can use to resolve any kind of phylogenetic relationship within primates, simply because it’s highly unlikely that they independently occurred at the same place on the DNA. Nevertheless they found one Alu SINE which humans and Gorillas had, but Chimpanzees didn’t.

But how polymorphic was the ancestral population of Gorillas, Chimpanzees and Humans on a phenotypic level?
This question is almost impossible to answer, at least for now. we have no Idea how the genotype relates to the phenotype as long as we aren’t able to close this gap we’re not able to make anything more than vague assumptions about this question. Looking at this whole story of Neanderthals and “Denisovans” interbreeding with modern humans (Greene et al., 2010; reich et al., 2010), which probably means that we have to broaden up our own species to include at least the Neanderthals if not even Homo heidelbergensis, I think we can see that a species can vary pretty much on a phenotypic level without interfering with reproduction.

This whole story definitely doesn’t make things easier since we now have two problems: Convergent evolution and incomplete lineage sorting. But I think we can at least try to differentiate things a little bit more. Although we have no Idea how much incomplete lineage sorting has affected the phenotype of the African great apes and humans we can’t rule out that it had no effect at all. Therefore it’s pretty much futile to make any assumptions about the “quality” of certain characters, since we have no idea to what extent those characters were affected by incomplete lineage sorting. As we could see, characters can be pretty much invulnerable to parallel evolution and still indicate wrong phylogenetic relationships (Salen et al. 2003).
Maybe we should instead explicitly look at each and every character, especially those which support the “correct” tree and test, if they share a common origin.


Collard M, & Wood B (2000). How reliable are human phylogenetic hypotheses? Proceedings of the National Academy of Sciences of the United States of America, 97 (9), 5003-6 PMID: 10781112
Green, R.,et al (2010). A Draft Sequence of the Neandertal Genome Science, 328 (5979), 710-722 DOI: 10.1126/science.1188021
Reich, D., et al. (2010). Genetic history of an archaic hominin group from Denisova Cave in Siberia Nature, 468 (7327), 1053-1060 DOI: 10.1038/nature09710
Salem, A. (2003). Alu elements and hominid phylogenetics Proceedings of the National Academy of Sciences, 100 (22), 12787-12791 DOI: 10.1073/pnas.2133766100
Satta Y, Klein J, Takahata N (2000). DNA archives and our nearest relative: the trichotomy problem revisited. Molecular phylogenetics and evolution, 14 (2), 259-75 PMID: 10679159
Strait DS, Grine FE (2004). Inferring hominoid and early hominid phylogeny using craniodental characters: the role of fossil taxa. Journal of human evolution, 47 (6), 399-452 PMID: 15566946
Wood, B., Harrison, T. (2011). The evolutionary context of the first hominins Nature, 470 (7334), 347-352 DOI: 10.1038/nature09709

P.S.: I just realised that used almost the exact same formulation for the title of this Post as I used for my last title. I'm deeply sorry for my lack of creativity. (11.10.2011)