GENEALOGY-DNA-L Archives

From: "Diana Gale Matthiesen" <>
Subject: RE: 12 marker match - rare v. common haplotypes
Date: Sat, 1 Apr 2006 21:11:27 -0500

I got a request off list to, among other things, expand on my answer, which I
did. And as long as I've written this up, I'll share this part of it with the
list, in case anyone is interested...

============================
All you can do with DNA test results is falsify hypotheses. In other words, a
DNA mis-match is proof that no connection exists, while a DNA match simply
leaves the door open to there being an actual match. This is true of *any* DNA
test. When the results do not falsify the hypothesis, you then weigh the
probabity that the positive result means a true connection. This is why
forensic geneticists, acting as expert witnesses in criminal trials, speak of
there being a one in 10,000 or one in a million, etc., chance that anyone other
than the defendant is the perpetrator. Still, the DNA has not proven the
defendant is THE perpetrator, it has only proven it extremely unlikely that it
is anyone else.

A 12/12 match between two people with a paper connection is supportive of the
connection being correct, but how much that 12/12 match supports the connection
DEPENDS ON HOW COMMON the 12-marker haplotype is. If the 12-marker haplotype is
extremely common (and, by definition, the most common haplotype is, well, THE
most common haplotype), then the match may only give modest support. If the
haplotype is rare, it may add a great deal of support. I can give you examples
of both extremes in my project:

COMMON HAPLOTYPE

http://dgmweb.net/genealogy/DNA/Straub/StraubDNA-results-HgR1b.shtml

If you look at the row labeled "Modal R1b1 Haplotype," that is THE most common
haplotype in western Europe (called the WAMH). Note that the first subject in
the table, STROOP #13668, matches the WAMH exactly, but if you look at his
lineage (scroll to the far right), his ancestor is a STRUBE, not a STRAUB, like
the next three.

If you then look at the three STRAUB's labeled "Bieringen Straub's," who have a
common ancestor on paper, you'll see that they, too, match the WAMH at
12-markers (except that one of the two cousins has had a recent mutation). If
we compare #34442 (the STRAUB who is native to Bieringen) and #37265 (the cousin
without the mutation), their 12/12 match drops first to 23/25, then to 32/37,
which raises a question as to whether one of them hasn't made a mistake in their
paper genealogy. No matter how certain you are that the paper genealogy is
correct, I would never stop short of 37 markers before making any declaration as
to how related two people are (beyond the fact that a strong mis-match means no
relationship).

RARE HAPLOTYPE

http://dgmweb.net/genealogy/DNA/Straub/StraubDNA-results-HgI.shtml

If you look at the four subjects at the top of the table labeled,
"I1a-Anglo-Saxon-5," you'll see that all four have a 12/12 match and have at
least a tentative, if not proven, paper connection to the same progenitor. This
haplotype is so rare that only these four people (compared to any online
database) have this haplotype; they don't even have any near matches (33/37 or
better). When compared at 25 markers, two match 25/25 and the others 24/25. At
37 markers, two match 37/37, one matches 36/37, and one matches 35/37. So, in
addition to the paper connections, two aspects of the DNA results strongly
support the connections: 1) their haplotype is unique and shared (so far) only
by them and 2) two of them have a 37/37 match and none is less than 35/37. At
the moment, anyone with their surname who tests only 12-markers and matches them
12/12 has a near certain (like 99.99%) probability of being related. In fact,
until someone outside the surname (or someone who descends on paper from a
different progenitor) turns up with this exact 12-marker haplotype, this
12-marker haplotype could be considered definitive for this progenitor.

These examples are why I say it matters *very much* whether the 12-marker
haplotype is rare or common as to how much significance you should attach to it.

A hypothetical parallel situation would be if two SMITHs bump into each other on
the street, compared to having two SNITZELHUMPERDINKS bump into each other. The
SMITHs would probably say, "Ho hum, ta ta," but the SNITZELHUMPERDINKS might be
overcome with excitement to think they might have found a long-lost cousin.
Having a rare haplotype or a common one is genetically parallel to having a rare
surname or a common one.

Hope this helps,
Diana

This thread:

• RE: 12 marker match - rare v. common haplotypes by "Diana Gale Matthiesen" <>