GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2010-12 > 1293659777


From: Vincent Vizachero <>
Subject: Re: [DNA] The death of paragroups
Date: Wed, 29 Dec 2010 16:56:17 -0500
References: <8FB365E1-2876-4664-AA70-3CCD4CEBB9CB@vizachero.com><000501cba78a$0850ce70$c2482dae@Ken1><003b01cba79b$5dbd2f00$19378d00$@dgmweb.net>
In-Reply-To: <003b01cba79b$5dbd2f00$19378d00$@dgmweb.net>


On Dec 29, 2010, at 4:00 PM, Diana Gale Matthiesen wrote:

> I share your puzzlement. It is not an "idea" that the Y-tree is
> "almost"
> entirely a set of bifurcations, it is a bifurcating tree.

There is no way to know that. I said "almost" purposefully: while it
is, indeed, likely that most nodes have only two extant descendants it
is likely that somewhere in the tree there is an exception that has
three or more. There is no law of nature that would prevent it.


> We simply have yet to
> determine the order of all the SNPs.

That's an odd statement to make, in that it intimates that we already
know what all the phylogenetically useful SNPs ARE and we just don't
know the ordering. Nothing could be further from the truth.


> Presumably, if we test enough people we
> will eventually know the ordering, in which case we will have a
> bifurcating
> tree, with a single SNP at each node . . .

Close, but no. For one thing, the SNPs are best thought of as sitting
on the branches and not at the nodes. For another, the number of
branches that will ultimately have just a single SNP is infinitesimal.

If you are talking about the tree representing living men, for most of
human history the branches are FAR too long to be occupied by a single
SNP. Most will have dozens, and some hundreds, of SNPs. Right now I
could give you a list of more than three dozen SNPs that occupy the
branch between R1b1b and R1b1b2 and which will never be broken up
using extant taxa.

Even if you were talking about some hypothetical tree, representing
all men who have ever lived, the chances of us finding useful DNA from
every one of those men is nil.

VV


This thread: