Archiver > GENEALOGY-DNA > 2008-09 > 1221431264

From: "Diana Gale Matthiesen" <>
Subject: Re: [DNA] Variability of DYS 570
Date: Sun, 14 Sep 2008 18:27:44 -0400
References: <>
In-Reply-To: <>

I appreciate the work you have done, David, but I'm taking an entirely different
approach to using these data, possibly because the majority of my members have
substantial paper pedigrees. In fact, if they don't have a pedigree, I "do"
their genealogy for them (i.e., just their patrilineal line), because I don't
want data in my projects without lineages. It's one of the perks of joining my

The fact is, I don't believe genetic distance, per se, is much use to a
genealogist. I have brothers with a genetic distance of one, sixth cousins with
a genetic distance of zero, and fifth cousins with a genetic distance of three.
In a genealogical time frame, genetic distance tells you very little beyond the
fact that small distances mean a relationship is possible and great distances
mean a relationship is impossible. Virtually all of my members (in my surname
projects) are interested in just two things: supporting their paper pedigree
and "crossing the pond." They are not interested in their deep ancestry and,
even if they are, they aren't using their surname project to study it.

My misconception, I now see, is that I underestimated the probability of a
parallel or back mutation in genealogical time, largely because my projects have
never had one before, so I never bothered calculating it before. I see now
that, while the probability is low, it's not nearly as remote as I had thought,
not when you have a large number of people testing 67 markers.

I think Didier's suggestion to test more markers is the answer in this case
because there are some markers in the "advanced" group that also divide this
family in two. I will be pressing this member -- and all my members who want a
fine-grained test of their pedigree -- to max out their markers. If this "odd"
individual matches his paper group on the advance markers, then we'll know he's
had a parallel mutation. If this individual matches the other group on his
advance markers, too, then we need to start looking for another immigrant from
this family to have come to America because the probability of having three or
four parallel mutations in nine or ten generations really is too remote to
consider. Or someone please tell me it is...


P.S. My browser balked at loading some of your pages, so I haven't seen them
all, but I do have a comment based on what I did see. If this family is R1b,
why are you testing only 37 markers? A GD of 5 at 37 markers can represent a
non-relationship, especially in R1b. And why are you assuming all of these
EWINGs have a common ancestor? It seems to me the reason you are having so many
back mutations and parallel mutations is that you are trying to connect
unrelated families. I would regroup your members based on their paper
connections, leaving the ones without a paper connection unassigned, and see how
many parallel and back mutations you have then. That is how I group my members,
with the result that this is the first time in four years of running six
projects that I've seen a parallel/back mutation. I really do think such
mutations should be rare, and definitely not *more* common than unique

> -----Original Message-----
> From: On Behalf Of David Ewing
> Sent: Sunday, September 14, 2008 10:07 AM
> To:
> Subject: Re: [DNA] Variability of DYS 570
> We have a very similar situation at DYS 391 in the Ewing
> Project. The Ewing
> modal is 11 at this marker, but a large subset of our
> participants have 10.
> Though not all of our participants have well worked-out conventional
> genealogies, among those that do, the DYS 391 branch marker
> is robust except
> for one individual, who appears in the "wrong" family.
> Perhaps I should say in this connection that we have 57
> individuals in a
> cluster within genetic distance 5/37 from the Ewing modal.
> Among these,
> there are a total of 72 mutations, of which 19 are back
> mutations (6 of
> these are at CDY), 35 are parallel mutations (9 of these at
> CDY), and only
> 18 are unique mutations. As may be obvious to some, the fact
> that only 30 of
> these individuals have well worked-out conventional
> genealogies may have
> resulted in my miscounting parallel mutations, but to the
> extent possible, I
> placed all these individuals on a hypothetical phylogeny
> diagram in a way to
> minimize back and parallel mutations while respecting conventional
> genealogies.
> If you would like to have a look at this diagram, it is available as a
> downloadable Excel file at
> k/Y-DNA_Phylogeny.html
> It is a large diagram and is virtually impossible to look at
> on the computer
> screen, so I recommend printing it out on at least a couple
> of letter-sized
> pages.
> David Ewing


This thread: