GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2010-02 > 1266862192


From: David Faux <>
Subject: Re: [DNA] Background relatedness and geneological relatedness
Date: Mon, 22 Feb 2010 10:09:52 -0800
References: <841343.79239.qm@web24503.mail.ird.yahoo.com>
In-Reply-To: <841343.79239.qm@web24503.mail.ird.yahoo.com>


Anders,

Since Decodeme introduced a new feature which allows customers to examine
fragment size from 250 Kb to 10 Mb, it has been quite a learning experience
as to who we match and don't match. When looking at the reference samples
(one from each of the 52 populations) it seems clear that most Europeans
will have multiple matches across the Eurasian world (and the Americas) at 3
Mb (no cM values are available - but one can look at SNP density via the
browser). The two that turn up time and time again are half way down the p
end of chromosome 3 and the centromere of chromosome 11. Some even retain
"relatedness" at the 5 Mb level. There seem to be odd inexplicable matches
with some of the reference samples. I match say an East Asian Oroquen on
chromosome 4 and a Maya on the X. It is amazing how quickly the matches
drop away when changing from 3 to 5 Mb, and then none are left when upping
it to 10 Mb - although I can see that occasionally a say 8 Mb match is
present when examining the area more closely with the browser.

Clearly it should be possible to remove the obvious noise from signal (chr 3
and 11 above), but there are some gray areas and it is difficult to know
what may be "meaningful" in some way. What of course is odd is that I have
a number of fourth cousins with whom I share zero (although they share with
others in the family), and yet I share as you say a match that is of a size
more characteristic of reasonably close relatives with an Adigey.

The matter is further discussed on the Decodeme Forum and DNA-Forums.

I am becoming more convinced that the evidence via 23andMe (only) is not at
all supportive of the paper trail which very clearly shows Native American
(I am beginning to wonder if this was a step mother to the children that is
described so clearly in a contemporary diary). Tim has reported that he has
one Asian block (1/256 Wyandot via genealogical sources), and that his
mother 6 and her first cousin 4 (each 1/128 Wyandot). Of the sundry members
of my family there are no Asian blocks that are not better explained via
Ukrainian ancestry. What has emerged though is small amounts of African in
those of one lineage of the family. I had forgotten that this branch
appears to descend from Eve (Classen) Pickert who Sir William Johnson termed
a mullato (she spoke perfect Mohawk and sold liquor to the them from her
home near Canajoharie). So if 23andMe can be trusted here to provide valid
information then no Native American, but African ancestry - which can be
woven into the genealogical record to reveal the true story (I hope).

David K. Faux.

On Mon, Feb 22, 2010 at 1:42 AM, Anders Pålsen <> wrote:

> David, list.
>
> I need to correct myself here as it seems I remembered wrong. My Pima
> X-match is above 6 Mb but the equivliant cM distance was only a little less
> than 2 cM. So the segment is in a high preservation area of the X. Still the
> segment is far past its predicted survival time using simple recombination
> rate theory.
>
> This emphasis the reoccuring problem of background relatedness vs
> geneological relatedness, where does the last former ends and later begins
> and is a segment match real or just noise? I have plenty information in my
> data to be able to form some kind of "theory" and some seems problematic for
> example David you have a 7 cM match with an Adigey at chromosome 2. I wonder
> if this individual would match you in 23andme's Relative Finder.
>
> Anders
>
>



This thread: