Molecular taxonomy: on way out?

[John Grehan]

>Date: Sun, 17 Jul 2005 00:08:53 -0400

>To: Richard.Zander at MOBOT.ORG, TAXACOM at LISTSERV.NHM.KU.EDU

At 04:26 PM 7/16/2005, Richard.Zander at MOBOT.ORG wrote:

 

>Molecular taxonomy will be old hat sooner than you think.

 

It will be only when morphologists no longer subordinate morphology to 

DNA sequences as a conditioned reflext

 

>And the leading

>expositor on Taxacom for a new and better way to evaluate evolutionary 

>patterns as they may be used in taxonomy in the future is . . . (wait 

>for it). . . John Grehan.

 

Hmm. I'm not sure what to make of that.

 

>In my opinion, John got to the right conclusion with metaphysical
premises.

 

Thanks, although I don't know what the right conclusion was

 

>We, rightly, disagreed with Grehan's "molecular drive" and other 

>generalizations as an alternative to modern molecular systematics.

 

Me and Queen Victoria. Molecular drive is part of 'modern' molecular
genetics.

 

>The problem is that some major morphological complexes are greatly 

>split by molecular systematics.

 

It would seem to be a problem only if one views DNA sequences as 

somehow the essence of phylogenetic relationship.

 

>Although statistically the molecular split is well supported, that 

>split statistically contravenes Dollo's Law that it is not to be 

>expected by chance alone that two taxa sharing many complex 

>morphological characters should separately exactly re-evolve (by 

>gradual accumulation of traits).

 

Phenetic characters (that are a combination of primitive and derived
states)that are statistically well supported are still just phenetic
characters. 

 

>selection on regulatory genes. It is the only explanation of such 

>homoplasy, it is not mystical, and it builds on Darwinian thinking.

 

Or the DNA sequence similarities occur because the characters used in
molecular analyses comprise plesiomorphic as well as derived states. The
a posteriori use of outgroups to noothing more than cladistic mimicry
that assumes that the outgroup 

will automatically define the real plesiomorphic states (and the same
criticism applies to morphological studies that use unordered character
states). This method is further problematic for DNA sequence analyses in
that it contradicts molecular theory that posits divergence of all
daughter lineages and yet at the same time the outgroup is supposed to
represent unchanged (primitive) character states!  

 

>The human-chimp relationship is statistically supported by molecular 

>studies as a sister group,

 

Yes - maybe statistically supported for DNA sequence characters, but 

not necessarily by morphological characters!

 

>but we share too many critical morphological traits with the 

>patristically more distant orangutan

 

Well maybe the orangutan is not patristically more distant. That's the 
whole point of the orangutan evidence Accepting 

the orangutan as phylogenetically more distant because of DNA sequences 

seems to contradict the assertion above that I'm an expositor of a 

better way (actually I don't know that I am - I just take the position 

that there is no evidence to necessitate morphological relationships 

being subordinated to DNA relationships when the two are incongruent 

and that when there is a lot of strong morphological evidence perhaps 

that indicates in specific instances that the DNA sequence 

relationships are not tracking phylogeny as well as the morphology).

 

>not to suspect some kind or degree

>of direct shared genetic past, a deep homology.

 

That seems to indicate that the uniquely shared morphological 

similarities are just somehow plesiomorphies because they are
incongruent with molecular trees.

 

>If demonstrably shared evolution is more important to systematics and 

>classification than tree-like segregation of gradually accumulated 

>changes in introns and junk DNA, and if evolution may not be recovered 

>in at least some salient cases by molecular systematics, then careful 

>examination of morphology is a major clue to evolutionary 

>relationships. So, dust off your 1970's phenetic analysis software!
Morphology will rule again!

 

I agree with the first sentence (which seems to contradict Zander's 

support for statistically well supported DNA sequence phylogenies over

morphology) but I do not see the necessity for the second sentence. 

Seeing morphology as a major clue to evolutionary relationships does 

not necessitate phenetic (non-cladistic) analysis where characters used
in the ingroup are not limited to derived states. Instead, I would use
the cladistic method of 

Rosa, Hennig, and Nelson where the characters analyzed are only those 

that are each individually shown to be present only in the ingroup.

 

If I've shown my ignorance in misunderstanding anything I'm sure I will 

be corrected.

 

John Grehan

 

 

Dr. John R. Grehan

Director of Science and Collections

Buffalo Museum of Science1020 Humboldt Parkway

Buffalo, NY 14211-1193

email: jgrehan at sciencebuff.org

Phone: (716) 896-5200 ext 372

 

Panbiogeography

http://www.sciencebuff.org/biogeography_and_evolutionary_biology.php

Ghost moth research

http://www.sciencebuff.org/systematics_and_evolution_of_hepialdiae.php

Human evolution and the great apes

http://www.sciencebuff.org/human_origin_and_the_great_apes.php