GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2007-10 > 1192149927


From: (John Chandler)
Subject: Re: [DNA] CODIS Markers and the Program STRUCTURE
Date: Thu, 11 Oct 2007 20:45:27 -0400 (EDT)
References: <147566.67370.qm@web50704.mail.re2.yahoo.com>
In-Reply-To: <147566.67370.qm@web50704.mail.re2.yahoo.com> (message from DavidFaux on Thu, 11 Oct 2007 09:30:51 -0700 (PDT))


David wrote:

> This is one of the few tests that can accurately address admixture.

Note that getting an assessment of admixture is very different from
making a list of likely affinities. In the latter case, you are
assuming there is NO admixture and simply trying to find the ONE
ethnic group that probably provided ALL of the detected alleles. In
the former case, you are assuming that you have sampled all of the
existing ethnic groups (and there had better be a SMALL number of
groups) and trying to find the percentage of the tested genome that
came from each group. Obviously, any part of the genome that is not
physically close to any tested marker is "invisible" to this
assessment, and therefore a meaningful admixture test simply cannot be
done with only 22 markers. 66 evenly spaced markers would seem to be
the bare minimum, and it would be better to have a lot more.

John Chandler


This thread: