Archiver > GENEALOGY-DNA > 2007-01 > 1170290311

From: "William Hurst" <>
Subject: Re: [DNA] New Ashkenazi DNA paper
Date: Wed, 31 Jan 2007 19:38:31 -0500
In-Reply-To: <>

Hi Ellen and all,

Comments below.

>I completely agree with you. It would have been far
>more revealing to provide a sub-clade comparison.
>Behar's mtDNA data indicates significant differences
>in mtDNA haplogroup and haplotype frequencies,
>particularly between Eastern European and Western
>European Ashkenazim. While it is fascinating to see
>population substructure between the Eastern European
>Ashkenazi communities, it would have been very helpful
>to also include a comparison with German Jews as well.
> This seems particularly important given the claim
>that Ashkenazim of both Central and Eastern Europe
>originated "exclusively from the Rhine basin."
>Has a study been done then between Ashkenazim mtDNA
>and that from the Rhine basin? Feder's report claims
>that it has, but cites no references or results.
>The problem with the Rhineland theory is that is too
>limited in scope. While there appears to a shared
>history between German Ashkenazim and Eastern European
>Ashkenazim, there are also significant differences
>that cannot simply be explained by genetic drift.
>There are haplogroups and haplotypes present in
>Eastern European Jews that are not found in German
>Jews. Thus, while there was probably an influx from
>Germany into Eastern Europe during the 14th century,
>that influx must either have merged into a
>pre-existing Eastern European Jewish community, or in
>the alternative, there must have been some gene flow
>from Eastern Europeans into a primarily Germanic
>Jewish group.

I just took another look at the K entries in MitoSearch for
Eastern Europe (EE), Germany and Poland. I only looked at
those with high-resolution or HVR1/2 results. For EE, I used
Russia, Ukraine, Belarus, Lithuania and Romania. Latvia and
Bulgaria had no hi-res results.

Of the 39 EE entries, 87% are in one of the three Ashkenazi
subclades, with 51% in K1a1b1a and 18% each in K1a9 and
K2a2a. Remember that overall, the three Ashkenazi subclades
represent only about 20% of K. The range is very narrow.
I'll try to use a little table, adding Germany and Poland:

Country--Total--Ashk %--K1a1b1a %--K1a9 %--K2a2a %


K1a1b1a is bi-modal; half from EE have 16223T, half don't,
with not much variance by country. All have 114T. One from
Ukraine has a pair of 309 insertions. No other differences.
All the K2a2a are modal. All the K1a9 have the added
16093C to the basic haplotype, so that's also modal.
Mapped out, it appears that K1a9 concentrated in the
southern part of EE, K2a2a in the nothern part, and K1a1b1a
speard out in an even manner. I created a map back in August
from counts of these subclades plus K1c and K1c2. See:

In Poland, about twice as many K1a1b1a have 16223T as not.
One without has 195C and 309.1C. One K1a9 does not have
1093C. The two K2a2a are modal.

In Germany, none of the three K1a1b1a has 16223T, but one
has 152C added. Both K1a9 have 16093C. The K2a2a is modal.

K1a9 and K2a2a are too young to have much variation anywhere.
K1a1b1a seems to have picked up 16223T after Germany. It has
slightly more added mutations in Germany and Poland than in EE.

As for the "parents" of the Ashkenazi subclades, Lithuania has
one apparent K2, which would actually be a "grandparent" of
K2a2a. Nothing else in EE; not one example of a non-Ashkenazi
K1a, which is about 60% of K. Poland has one K2a parent and
a couple of non-Ashkenazi K1a. Also there is a K1a with 195C
and without 524 insertions; I call that "pre-K1a9." Germany has
two each K1a and K2, but no direct "parents." Germany is the
true crossroads for K, having in addition substantial numbers of
K1b and K1c subclades. Poland is only missing K1b. I didn't look
for this little survey, but the Askenazi "parents" seem to have
moved mostly to the British Isles and Western Europe generally.
They certainly didn't go to EE.

>Bill, regarding haplogroup K, I'm not sure I
>understand the statement "I found 84% of EE in the
>Ashkenazi subclades, but only 65% from Poland." Could
>you clarify this statement for me?

Maybe I clarified this above, with slightly different figures
from new results: % of each area in the Ashkenazi subclades.

>Also, are you stating that you believe Ashkenazi
>subclade K2a2a possibly derived from Polish K2a2a?

Nope, I don't have a way to determine non-Ashkenazi K2a2a,
if there is such a thing.

>It is my suspicion, based on Behar's own tables, that
>Ashkenazi K2 is of European origin, although Behar
>states that all four of his "founding mothers,"
>including K2, are of ancestral Middle Eastern (ie,
>ancient Israelite) origin. This seems extremely
>unlikely to me.

If I am reading Behar's 2006 paper correctly, only K2a2a is
Ashkenazi; K2 and K2a are not. I would think K2a2a is too
young to have traveled from the Middle East to the Rhine
Valley to EE. But there are a lot of mutations, control- and
coding-region, between K2 and K2a2a.

>You further note that Poland is rich in the two
>Ashkenazi sub-clades K1a1b1a and K1a9. Again, to what
>do you attribute this to - gene flow from Ashkenazim
>into the non-Jewish Polish communities, or vise versa
>(or neither)?

Ashkenazi stopping along the way to EE. Austria also has
high percentages of Ashkenazi subclades, but I didn't look
at it today.

>...Behar does note that "K1a1b1a sister
>lineages, which share with it a common ancestry at the
>internal nodal level of subclade K1a1b1, can be found
>in Portugal, Italy, France, Morocco, and Tunisia."
>How this supports his contention of a Middle Eastern
>origin for Ashkenazi K1a1b1a and no gene flow from
>Europeans (or at least North Africa) is puzzling to

Behar's study was of Jewish mtDNA and he concentrated
on examples from Mediterranean countries. I recently
looked for K1a1b1a's "parents, grandparents and siblings,"
K1a1b1, K1a1, etc., and found them mostly in the British
Isles and Western Europe. Not many Tunisians in
MitoSearch. Not a lot of Irish DNA in what Behar looked at,
I don't think. For example, he found one example of the
16048A cluster - in a non-Jewish Moroccan. FTDNA has
at least 54 in its database, highly concentrated in those
with Irish ancestry. Before I put my K1c2 (Irish or British,
I assume) on GenBank, there was only the one he
found - in a Jewish Moroccan.

>I am fascinated that you have studied the parent
>sub-clades of Ashkenazi K and determined them to be
>Western European in origin. Am I interpreting your
>statement correctly? If so, the assertion is in
>disagreement with Behar, who states "the observed
>global pattern of distribution renders very unlikely
>the possibility that the four aforementioned founder
>lineages entered the Ashkenazi mtDNA pool via gene
>flow from a European host population."

I only say that the "parents" of the Ashkenazi subclades
are in Western Europe now, not Eastern Europe. Nothing
about origins of the subclades. When I use the word
origins, I usually mean the origins listed by the MitoSearch

>In support of this contention, Behar claims he didn't
>find similar HVRI mutations between Ashkenazi and
>British, Irish, German, French, or Italian subjects.
>He also claims these lineages are present in exile
>Jews who claim an origin from Spain before 1492. He
>had some other poorly supported contentions I won't
>bother going into, but I'd appreciate hearing your
>response to Behar's theory.

I'll let you argue with Behar! I don't have any way to
identify exile Jews from Spain. I will say that I think the
K1c2 and 16048A from Morocco were probably
descended from Portuguese or Spanish Jews who were
expelled in the late 15th century. There are historic
connections between these countries and Britain
which could have easily brought the mtDNA that far.
I see no way that the Irish and British K's are descended
from these Moroccans.

Bill Hurst

This thread: