[Taxacom] ad hoc manipulation

[Ken Kinman]

Hi Richard,
     I agree that some of the early studies on ape molecular sequences were 
sort of iffy and only moderately convincing, and your "coin flip" criticism 
is certainly valid where they were only looking at single base changes or 
perhaps even single base insertions/deletions at given loci.  However, we 
now have more complex sequence data analyses involving much longer sequence 
insertions (such as SINES) which are far more convincing since they are a 
lot less vulnerable to homoplasy.

     I would especially recommend a careful reading of Salem, et al., 2003  
("Alu elements and hominid phylogenetics"; Proc. Natl. Acad. Sci., 
100:12787-12791).  They analyzed the sequences of many such Alu SINES, 
including not only African apes, but orangutans, hylobatids, and ape 
outgroups.  It VERY strongly supported the phylogeny:  hylobatids 
(orangutans (gorilla (chimpanzees) man))).

     Such data indicates to me that there is still a reasonable chance that 
gorillas and chimpanzees actually clade together to the exclusion of man, 
but the chance that orangutans and humans clade together to the exclusion of 
chimps and gorillas is so remote that I find it unreasonable to pursue it 
any further unless the supporters of that hypothesis (such as Grehan) can 
counter such evidence with even a modicum of molecular evidence of their 
own.  They have nothing molecular to support their claim, and their 
morphological data has a "non-independent" cloud hanging over it that does 
not seem to hang over the SINE data.

     Extraordinary claims require extraordinary data to back them up, and 
the morphological data presented by Grehan is FAR from extraordinary in my 
opinion, and the molecular data for his hominid-orangutan clade is pretty 
much non-existent.

      Therefore, I think the data continues to support the orangutan 
splitting off first, followed by a paraphyletic stem-lineage that gave rise 
to the African ape clade (gorillas, chimps, and humans).  Whether gorillas 
or humans then split off first from the crown-group African clade is 
something that should be pursued more vigorously than the far less likely 
possibility that an exclusive human-orangutan paraphyletic stem-lineage ever 
existed.  Some molecular (single base) evidence may be weak, but it is 
corroborated by very strong SINE data.  The morphological data presented by 
Grehan continues to underwhelm me, while I suspect even stronger molecular 
data will corroborate the SINE results.  I certainly agree that strictly 
cladistic thinking may well "strait-jacket" our thinking in the phylogeny of 
some taxa, but I do NOT believe that great ape phylogeny is one of them.
       ----Ken Kinman
********************************
>From: "Richard Zander" <Richard.Zander at mobot.org>
>To: "John Grehan" <jgrehan at sciencebuff.org>,"TAXACOM" 
><taxacom at mailman.nhm.ku.edu>
>Subject: Re: [Taxacom] ad hoc manipulation
>Date: Wed, 5 Sep 2007 16:26:03 -0500
>
>Given that the true answer is probably "no one will ever know" I expect 
>this problem to turn up again and again. I think John Grehan has been 
>saying that the number of shared traits among man and orang is most 
>parsimoniously explained by direct ancestry. Well, that's a good 
>hypothesis.
>
>Ken Kinman replied that one needs some molecular support to even consider 
>the morphological data as compelling. Well, I don't know. There is no 
>reason the molecular data must track evolution of expressed traits. First, 
>the data are statistically very different. Molecular data are like flips of 
>a coin that indicate, en mass, whether a coin is (phylogenetically) loaded. 
>Morphological data are measurements of the coin to see if it is loaded. 
>Particular combinations of morphological data are what the molecular data 
>are supposed to be tracking.
>
>Molecular data on (man chimp) gorilla have been summarized by Satta et al., 
>and many loci support either (chimp gorilla) man or (man gorilla) chimp. 
>Has anyone checked how many loci support variants on where orang fits into 
>a tree? Could be some different gene histories there confounding things. On 
>the whole, it seems to me that molecular data significantly support ((man 
>chimp) gorilla) orang, but I've not seen a summary of all the data, just 
>some few studies.
>
>I've also pointed out in the past that a few genes coding for man/orang 
>morphological traits may have been silenced, then unsilenced, which would 
>explain why molecular and morphological results conflict.
>
>Also, note that the idea of monophyly has strait-jacketted evolutionary 
>thinking in that a surviving ancestor similar to man-orang (expressed 
>traits stabilized by selection) may well have thrown up an orang, then a 
>gorilla, then a chimp, then anagenetically ended up man. The ancestor is a 
>paraphyletic man-orang, and only two of the three surviving daughter taxa 
>are, through selection, different. Molecular traits continue to change in 
>the tree, but not the morphological traits.
>
>Thus, we have two possible ways that plain old dichotomous morphology 
>(expressed traits) trees may well differ from molecular phylogenies. 
>Demonstrating details of expressed trait evolution (the important kind) may 
>take fossils, or some new method, or may never be capable of the kind of 
>resolution apparently achievable in molecular analysis of gene trees.
>
>________________________________
>

_________________________________________________________________
Discover sweet stuff waiting for you at the Messenger Cafe.  Claim your 
treat today! 
http://www.cafemessenger.com/info/info_sweetstuff.html?ocid=TXT_TAGHM_SeptHMtagline2