More on human-orangutan analysis

[Will Fischer]

There are some points missing from John Grehan's recent post:

1. The analysis of craniodental characters by Collard and Wood(*)
doesn't strongly support a Homo-Pongo clade: the most parsimonious tree
for the full dataset was (Hylobates,(Homo,(Pan,(Gorilla,Pongo)))) --
the (Homo,Pongo) clade was from a bootstrap consensus tree only.

2.  Regardless of the conclusions of the paper, the point to the
exercise (which Dr. Grehan apparently doesn't recognize) is to identify
the discrepancy between the craniodental characters and the molecular
characters -- one of the data sets must be giving the wrong answer
(which one, of course, remaining an open question).

3.  However, and finally, the analyses of soft tissue characters (both
in 2000(+) and 2002(++)) got the same tree as the molecular data.

So, lets summarize some results:

        molecular data:
                Homo,Pan

        soft tissue data
                Homo,Pan

        craniodental data
                Homo,Pongo
                        OR
                Homo,(Pan,(Gorilla,Pongo))

        "conventional hard tissue data" [unreferenced in (++), superset of
craniodental]
                Pongo,(Homo,(Gorilla,Pan))

If a majority of the morphological characters supported a Homo-Pongo
clade, you could make an argument for it.  Since the hard-tissue data
are ambiguous, while the soft-tissue data and the molecular data are
congruent, it's hard to be enthusiastic about the unconventional
hypothesis, or your advocacy for it.

Since this all started in the context of a museum exhibit, why not
present (a BALANCED selection of) all the data:  sections of sequence
alignment (with differences highlighted), and a series of pictures of
hard- and soft-tissue character states.  Get the folks who attend the
exhibit to map the character states on the trees;  if you provide
"answers", make them be of the form:
        IF TREE A IS TRUE,
        CHARACTERS w-x are convergent,
        CHARACTERS y-z are historical;

        IF TREE B is TRUE, etc.

With five taxa, there are only 15 unrooted trees (still maybe too many
to put them all up on the display).  But you can show how the
characters group on some of them, and then you can get folks thinking
about what evolutionary forces might cause some characters to converge.
  For it to be fair, though (and other people on the list have mentioned
it), you'll have to choose the characters on display proportionally
according to the tree they support.


-- References --
* Collard M, Wood B. (2000)
   Proc Natl Acad Sci U S A. Apr 25;97(9):5003-6.
   How reliable are human phylogenetic hypotheses?

+       Gibbs S, Collard M, Wood B. (2000)
        Proc Natl Acad Sci U S A. 2000 Sep 26;97(20):11130-2.
        Soft-tissue characters in higher primate phylogenetics.

++ Gibbs S, Collard M, Wood B. (2002)
    J Anat. 2002 Jan;200(Pt 1):3-49.
    Soft-tissue anatomy of the extant hominoids:
    a review and phylogenetic analysis.


On Wednesday, June 25, 2003, at 12:09  PM, John Grehan
<jgrehan at SCIENCEBUFF.ORG> wrote:
> I've belatedly realized that there is one published cladistic analysis
> of
> morphological characters that supported the Human-orangutan and
> chimp-gorilla clades. This was a study by Collard and Wood (2000) in
> Proceedings of the National Academy of Sciences (USA). Sounds good you
> might think? Not according to Collard and Wood. They had already
> decided
> that the molecular phylogeny is the only truth and any morphological
> studies not conforming to the molecular tree is wrong and the
> characters
> are uninformative.
>
> The irony of this study is that the 'method' employed by these authors
> invalidated their enterprise from the beginning since they started
> with the
> premise that morphology is, of itself, uninformative. So they wasted
> not
> only their own time and money, but that of the institutions from which
> they
> accessed specimens and resources of the Natural Environmental Research
> Council and Wellcome Trust, Leverhulme Trust and Henry Luce Foundation
> that
> provided financial support for what appears to be, from the
> perspective of
> the authors, a pointless exercise.
_____________________________________________________________
Will Fischer                       wfischer at uts.cc.utexas.edu

University of Texas at Austin      Lab Ph.: 512-232-7114
Integrative Biology                Lab Fax: 512-471-3878
1 University Station C0930
Austin, TX 78712-0253