GENEALOGY-DNA-L Archives
Archiver > GENEALOGY-DNA > 2004-09 > 1094417463
From: (John Chandler)
Subject: Re: [DNA] DNAPrint test -- cannot reliably distinguish between East Asian vs. Native American markers
Date: Sun, 5 Sep 2004 16:51:11 -0400 (EDT)
References: <410-22004905201822890@earthlink.net>
In-Reply-To: <410-22004905201822890@earthlink.net> (bobmclaren@earthlink.net)
Bob wrote:
> I'm also puzzled as to why Charles results do show EA and his
> brother's show NA, and this is consistant for both versions of DNAPrint.
That's the key to the whole business. What the test measures is not an
abstract ethnicity, but rather a specific, limited set of markers that
happen to be somewhat more frequent in one population than in another.
Almost all of these markers are found "naturally" in significant numbers
in *every* population. Charles and his brother each must have more than
the average number of "seemingly non-European" markers, and it happens
that Charles' set is a better match with Amerindian while his brother's
is a better match with Asian. There are few enough markers in the test
that ordinary fluctuations can have big consequences in the results.
> In any case, I agree with David. Charles has exposed an apparent flaw of
> monumental significance.
The flaw is simply a lack of precision, and it was implicit from the
very beginning in the large spread of the confidence contours. It was
inevitable that some such cases would show up when the test went to
mass marketing. Charles deserves credit for resisting the glib
explanations of his results and trying to find out more about the
workings of the test.
John Chandler
This thread: