[Taxacom] Congruence of gene trees

Sat Aug 15 17:13:56 CDT 2009

Hi Richard,

My replies to your points are given below.

Wellington, New Zealand.

My papers on biogeography are at: http://tiny.cc/RiUE0

--- On Sun, 8/16/09, Richard Zander <Richard.Zander at mobot.org> wrote:

From: Richard Zander <Richard.Zander at mobot.org>
Subject: RE: [Taxacom] Congruence of gene trees
To: "Michael Heads" <michael.heads at yahoo.com>
Cc: taxacom at mailman.nhm.ku.edu
Date: Sunday, August 16, 2009, 4:26 AM
Michael:

Interesting. So you are saying that an ancestor that is morphologically polymorphic gives rise to multiple descendants that reflect those polymorphisms variously. AND are you also saying that an ancestor that is molecularly polymorphic also gives rise to multiple descendants with various distributions of those molecular polymorphisms? 

MH: Yes. 

THUS are you also saying that any group of, say, three exemplars that are supported in three different ways by different gene lineages are all derived from that same ancestor. 

MH: Yes.

SO if there is evidence that there is differential lineage sorting, then molecular analysis cannot be used to distinguish a correct species tree for the taxa involved because there is no one correct species tree and all the taxa are derived from a joint ancestor with polymorphic genes? 

MH: You can still derive a single 'best tree' ('correct species tree') if you want, using whatever algorithm you favor. But if ILS occurs a tree may not be an accurate representation of phylogeny. 

So if John Grehan can find differential lineage sorting with well-supported but different gene trees for configurations of orang, man, chimp and gorilla, then the molecular evidence cannot be used to disprove or decide against a morphological analysis of relationships? Are you implying this? 

MH: It depends on the algorithm you use. But one character tree doesn't 'disprove' another. They may all be correct representations of phylogeny. Taking a 'total evidence' approach often obscures very interesting biogeographic patterns, e.g. one gene may show Africa - America, another may show Africa - Asia. Both may be of critical importance for evolution, dating etc. but only one can appear in a 'best tree'.  

Are you saying that if there are many gene trees supporting one tree and only are few supporting other configurations, then we should not choose the configuration (topology) of the one supported by the most or the overwhelming number of gene trees from different genes?

MH: No. If you think that more = better, common = good, then you can choose the most common one, if you feel you need to choose just one. Or you could use another algorithm. I was trying to explain how incomplete lineage sorting worked, not how taxonomy should deal with it. 

This is NOT following dogma, Michael! 

MH: Can you supply a reference? All this is pretty orthodox as far as I'm aware but ideas on the subject are evolving rapidly. The point is that serious incongruence ('rampant parallelism') is very common, e.g. in passerines, angiosperms, and probably most groups. One point on terminology: since about 2005 you'll find many more references to this key topic under 'incomplete lineage sorting' than under 'differential lineage sorting' - 1200 vs. 40 on the best bioinformatics website (Google scholar).
  The different gene trees that are incongruent with a 'best tree' are not 'wrong' as you suggested in your first letter. They reflect ancient variation in the ancestor and are of great interest for evolutionists, biogeographers and phylogeneticists. Incongruence was traditionally written off as (phylogenetically) meaningless noise superimposed on the true, ideal, neatly hierarchical phylogeny. ILS starts to suggest that phylogeny itself may not be neatly hierarchical. A taxonomy is just a rough summary of the phylogeny - it doesn't actually reflect it fully. Phylogeny as a neatly branching tree is just a metaphor, albeit a very old and very respectable one, and very useful for filing specimens and discussing things.     
  ILS is predicted to be much more likely in an ancestor that is (a) widespread
and (b) evolving rapidly. Both of these are accepted as standard characteristics of ancestors in vicariance theory.

Doubtless one of the excellent phylogeneticists among us will explain why my take (above) on your message is totally wrong.

R.

Wellington, New Zealand.

My papers on biogeography are at: http://tiny.cc/RiUE0

--- On Sun, 8/16/09, Richard Zander <Richard.Zander at mobot.org> wrote:

From: Richard Zander <Richard.Zander at mobot.org>
Subject: RE: [Taxacom] Congruence of gene trees
To: "Michael Heads" <michael.heads at yahoo.com>
Cc: taxacom at mailman.nhm.ku.edu
Date: Sunday, August 16, 2009, 4:26 AM

Michael:

Interesting. So you are saying that an ancestor that is morphologically polymorphic gives rise to multiple descendants that reflect those polymorphisms variously. AND are you also saying that an ancestor that is molecularly polymorphic also gives rise to multiple descendants with various distributions of those molecular polymorphisms? 

THUS are you also saying that any group of, say, three exemplars that are supported in three different ways by different gene lineages are all derived from that same ancestor. SO if there is evidence that there is differential lineage sorting, then molecular analysis cannot be used to distinguish a correct species tree for the taxa involved because there is no one correct species tree and all the taxa are derived from a joint ancestor with polymorphic genes? 

So if John Grehan can find differential lineage sorting with well-supported but different gene trees for configurations of orang, man, chimp and gorilla, then the molecular evidence cannot be used to disprove or decide against a morphological analysis of relationships? 

Are you implying this? Are you saying that if there are many gene trees supporting one tree and only are few supporting other configurations, then we should not choose the configuration (topology) of the one supported by the most or the overwhelming number of gene trees from different genes?

This is NOT following dogma, Michael! Doubtless one of the excellent phylogeneticists among us will explain why my take (above) on your message is totally wrong.

R.

*****************************
Richard H. Zander 
Voice: 314-577-0276
Missouri Botanical Garden
PO Box 299
St. Louis , MO 63166-0299 USA
richard.zander at mobot.org
Web sites: http://www.mobot.org/plantscience/resbot/
and http://www.mobot.org/plantscience/bfna/bfnamenu.htm
Non-post deliveries to:
Missouri Botanical Garden , 4344 Shaw Blvd. , St. Louis , MO 63110
*****************************

From: Michael Heads [mailto:michael.heads at yahoo.com] 
Sent: Friday, August 14, 2009 7:11 PM
To: Richard Zander
Cc: taxacom at mailman.nhm.ku.edu
Subject: Re: [Taxacom] Congruence of gene trees

Hi Richard,

Ancestors may be polymorphic (cf. Heads, M. 1985. On the nature of ancestors. Systematic Zoology 34: 205-215.). If lineage sorting in the ancestor is incomplete there is no 'clean slate' (pure monomorphism) in the ancestor before a new round of differentiation begins. The old polymorphism may be inherited and may have a spatial distribution quite different from the new break. The end result of incomplete lineage sorting in the ancestor - the incongruence - is similar to that of hybridism, but ILS occurred before the differentiation of the new subclades, hybridism after.  ILS explains incongruence in any character tree, molecular or morphological. See also: Heads, M. 2009.   Darwin ’s changing ideas on evolution: from centres of origin and teleology to vicariance and incomplete lineage sorting. Journal of Biogeography 36: 1018-1026.

Michael Heads

Wellington , New Zealand .

My papers on biogeography are at: http://tiny.cc/RiUE0

--- On Sat, 8/15/09, Richard Zander <Richard.Zander at mobot.org> wrote:

From: Richard Zander <Richard.Zander at mobot.org>
Subject: [Taxacom] Congruence of gene trees
To: "Jason Mate" <jfmate at hotmail.com>, "Taxacom" <taxacom at mailman.nhm.ku.edu>
Date: Saturday, August 15, 2009, 11:18 AM

Disparate congruence of results of different genes (sequences) between
the sequences has long been known. Many gene trees have apparently
different histories (differential lineage sorting). Some trios (like
man, chimp, gorilla) have most genes supporting one tree (man, chimp
terminal) and about equal but fewer numbers of genes supporting two
alternative full resolutions (man, gorilla terminal) and (chimp, gorilla
terminal). All or many gene trees may be well supported, which supports
their discrepant histories. But then the distributions become governed
by small sample statistics (requiring assumed distributions) in most
cases but a few.

The explanation is fairly well-known, but the functional basis as it
affects statistics is not understood at least by me. Okay, the null is
not that all three trees are equal in probability, but instead the null
is that only one tree is possible given shared ancestry. The null is
that some process forces some gene trees into wrong configurations, and
this process is equiprobable in the two alternative wrong
configurations. We then assume that the two least common configurations
are probably the wrong ones. I think that works okay. 

So what is the process that makes some trees wrong? Why is it
equiprobable for all possible trees? Is it "oh, I'm late homogenizing in
the population" and two polymorphisms are fixed during speciation
instead of one? Or some selection might be involved such that
equiprobable wrong configurations are not to be expected?

Someone enlighten us?

*****************************
Richard H. Zander 
Voice: 314-577-0276
Missouri Botanical Garden
PO Box 299
St. Louis , MO 63166-0299 USA
richard.zander at mobot.org
Web sites: http://www.mobot.org/plantscience/resbot/
and http://www.mobot.org/plantscience/bfna/bfnamenu.htm
Non-post deliveries to:
Missouri Botanical Garden , 4344 Shaw Blvd. , St. Louis , MO 63110
*****************************

-----Original Message-----
From: taxacom-bounces at mailman.nhm.ku.edu
[mailto:taxacom-bounces at mailman.nhm.ku.edu] On Behalf Of Jason Mate
Sent: Friday, August 14, 2009 5:31 PM
To: Taxacom
Subject: Re: [Taxacom] Molecules vs Morphology
I am somewhat confused by this. If evolution does not act on every
aspect of the organism equally all the time and lineages are independent
why would you expect congruence throughout? You can  have local
incongruence between any two particular datasets simply out of chance or
because the gene-tree of one molecular dataset is different from the
other due to the speciation process. It is the average of many datasets
that you want as your best-available-hypothesis.

_______________________________________________

Taxacom Mailing List
Taxacom at mailman.nhm.ku.edu
http://mailman.nhm.ku.edu/mailman/listinfo/taxacom

The Taxacom archive going back to 1992 may be searched with either of these methods:

(1) http://taxacom.markmail.org

Or (2) a Google search specified as:  site:mailman.nhm.ku.edu/pipermail/taxacom  your search terms here