Archiver > GENEALOGY-DNA > 2008-04 > 1209231932

From: "Robert Suydam" <>
Subject: Re: [DNA] Tale of Two Brothers
Date: Sat, 26 Apr 2008 12:45:32 -0500
References: <009a01c8a7b1$b0c5eb70$6400a8c0@Ken1><><00de01c8a7c1$27719120$6400a8c0@Ken1>
In-Reply-To: <00de01c8a7c1$27719120$6400a8c0@Ken1>

Any gene tree, I believe.

"Another constraint on inference from an evolutionary tree is the time
it takes all the current genetic variants to coalesce to a common
ancestral DNA molecule. There is no possibility of associations
between genetic variants and geography once this coalscence has
occurred as there is not longer any genetic variation. Coalescence
time therefore places an absolute limit on how far back in time
inferences can be made. Indeed, the theory of coalescence implies that
the first temporal half of the evolutionary tree of a DNA regions is
expected to consist of only two DNA lineages -- usually far too little
variation to contain significant phylogeographic information."

In this case, Templeton is discussing mtDNA:

"The time depth of the mtDNA tree has been estimated to be around
200-250 KY, implying that mtDNA is informative about human evolution
only to about 100-125 KYA. Given that the out of Africa expansion in
the replacement model is estimated at around 100 KYA, this event is at
the temporal extreme of informative inference from mtDNA. This in turn
raises serious logical flaw in using mtDNA to test the replacement

Just a "rule of thumb", but it does make some sense to me, and your
original post brought it to mind. I don't think Templeton is claiming
the entire most recent half of the tree will necessarily be highly
informative, only that the first half probably won't be.

On Sat, Apr 26, 2008 at 12:15 PM, Ken Nordtvedt <> wrote:
> Why don't you translate that into more specifics, if you wish. Does
> "region" refer to a region of the tree; of the y and mt dna trees or
> autosomal haplotype trees or both?

This thread: