GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2002-11 > 1038102017


From: "pwreed" <>
Subject: Re: [DNA] AncestrybyDNA test. Who's it for?
Date: Sun, 24 Nov 2002 01:40:17 -0000
References: <5.1.0.14.0.20021123173948.047e0da8@incoming.verizon.net>


Hi, I'm a scientist with loads of experience
in genetic marker (STRs and SNPs)
testing for medical research (not for genealogy
tho' but the principles are very similar).
I'm a keen family historian and hope I
can learn from as well as contribute to the list.

This is a relatively minor point about the AncestrybyDNA test,
and probably wont be a surprise to many of you,
but I suspect the AncestrybyDNA test has little value
to anyone with knowledge of their ancestry back at
least four generations and NO SUSPICION of admixture.
I suspect the test is unlikely to reliably determine any
unsuspected admixture from much before that time.

Put another way, the percentages of admixture
that can be reliably measured have to be large
(>> 5%) and therefore either very recent or very prevalent in the
general ancestral population.

I say its a relatively minor point because of course
many people know or suspect they have
recent admixed ancestry, and I guess
this is primarily who the product is aimed at.
Whether I think the test is of value for them/you/me/us
I'll address elsewhere.

Here's the reasoning for my main point, for those that want it:
The DNA that AncestrybyDNA tests is the
most common type, which you inherit 50%
from your father and 50% from your mother.
On average each grandparent contributed 25%
of your DNA, each G-grandparent 12.5%
and each GG-grandparent 6.25% and so on
up the generations. To use the example from
the AncestrybyDNA manual If you go back
10 Generations (or ca 1800), each ancestor
at this level has contributed only 0.1% to your
genetic makeup (excluding the Y and mtDNA).
By their own admission their ancestry values of
5% or less are unreliable (this figure is very
ballpark, so lets say its close enough to 6.25%)
which in turn means that on average their test
is insufficient to reliably detect the contribution of any
individual GG-grandparent (assuming only one GG grandparent
contributes to the admixture).

Peter Reed


This thread: