GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2002-08 > 1030371599


From: "Orin R. Wells" <>
Subject: Re: [DNA] How specific is 12 marker test
Date: Mon, 26 Aug 2002 07:20:00 -0800
References: <3D6A0F4E.A0BFE4FC@erols.com><003801c24b0f$2395aec0$59383942@nc.rr.com>
In-Reply-To: <5.1.0.14.0.20020826073824.056a1148@mail.verizon.net>


At 07:53 AM 8/26/02 -0400, John S Walden wrote:

>>Look at it this way The 12 Marker test has about a 1000 to 1 chance of
just being a random not related match between two people. But the case you
have is not two people it is 10,000 people.

I would say, given that FamilyTreeDNA can't have more than about 5000
samples in their data base, that in this case he matched 24 of them. That
was 1 in 200 (counting William himself). It could be even lower if they
have included mtDNA samples are included in this database. So I question
the 1 in 1000 number.

>You should EXPECT to get 10 matches. More if you have common numbers and
>fewer if you have rarer numbers.

What are common and what are rare? Are they rare simply because not too
many people with a given marker number have submitted samples to the FTDNA
database? Or are they common because they all possess some quirky gene
that perdisposes them to being interested in their ancestry.


>For any given set of 12 marker numbers (unless they are Really Really rare)
>will have 100,000 other men in the USA with those numbers.

Or maybe as many as 1 million. I don't think anyone really knows. If you
are looking at a fairly large surname study and you don't match that data
on a 12 marker test assuming you know that is the surname of interest, I do
agree that you can probably eliminate the person from belonging to one of
the known families. However, if you are participating in such a study, you
probably are going to need more to either define where you do fit or to
establish the pattern against which to match still others of the surname.



Orin R. Wells
Wells Family Research Association
P. O. Box 5427
Kent, Washington 98064-5427
<>
http://www.rootsweb.com/~wellsfam/wfrahome.html
Subscribe to the "Wells-L" list on RootsWeb

This thread: