False Positives? #dna


David Goldman
 

Hello, everyone at the DNA Group. I was notified of a new "extremely high"
probability of a 3rd or 4th cousin match >from Ancestry with 102 centimorgans
on 7 segments. However, when I was in touch with that match we shared all
the family names we could think of and none of them were the same. The
"match" would mean that our grandparents or great-grandparents were first
cousins. My known last names go back to my great-great-grandparents, and
even a couple more beyond that. So given this "false positive" I would have
to strike this up to the fact of high endogamy among Ashkenazi Jews, such
that perhaps we had ancestors several hundred years ago who were the same.
But for purposes of genealogy for most of us, such speculation is probably
rather meaningless.

David Goldman
NYC


David Brostoff
 

On Aug 4, 2018, at 7:09 PM, David E Goldman lugman@verizon.net wrote:

My known last names go back to my great-great-grandparents, and
even a couple more beyond that.
Have you and your match identified all your great-great-aunts and their
married names?

This is usually the limiting factor.

David Brostoff


Raina Accardi <raina.accardi@...>
 

Hi David,

Its hard to tell >from the Ancestry relationship estimate exactly what is
going on but an extremely high probability and that much shared DNA is
not a false positive. You two are definitely cousins of some kind. If
you can convince the match to upload to Gedmatch you will be able to get
a better estimate of the relationship by looking at the size of the
individual segments (larger sizes will indicate closer relationships) and
see if there are a lot of small segments (indicates endogamy). A 4th
cousin would share a 3x great grandparent so you may be looking at
roughly 150 years ago in the 1860s. Can you find any common locations
between your matches ancestors and yours at that time? If so, you might
be able to infer who the common ancestors are even if their surnames are
different. Many Jews took different surnames even among the same family,
some after the mother, some after the father, some made up new ones.
Another course is to look at the family trees of "shared" matches. Are
there any names and locations that connect to either your or your
match's tree? If so, that may be a line of further research.

Raina Accardi
Saugerties, NY

From: "David Goldman" <lugman@verizon.net>
Date: Sat, 4 Aug 2018 22:09:59 -0400

Hello, everyone at the DNA Group. I was notified of a new "extremely
high" probability of a 3rd or 4th cousin match >from Ancestry with 102
centimorgans on 7 segments. However, when I was in touch with that all
match we shared the family names we could think of and none of them . The
were the same "match" would mean that our grandparents or great- first
grandparents were cousins. My known last names go back to my great-great-
grandparents, and even a couple more beyond that. So given this "false
positive" I would haveto strike this up to the fact of high endogamy among
Ashkenazi Jews, such that perhaps we had ancestors several hundred years
ago who were the same. But for purposes of genealogy for most of us, such
speculation is probably rather meaningless.


Michael Good
 

Hi David,

In my experience a 102 cM match across 7 segments at Ancestry is
unlikely to be a false positive, but more likely a 3rd to 4th cousin
as Ancestry predicts.

Do you and your match really know all the descendants of all your
great-great-grandparents? I sure don't. Heck, I don't even know all my
great-great-grandparents yet.

If you don't know all your great-grandparents siblings and all their
descendants, then you are going to find names that you don't recognize
on your DNA matches >from when female descendants marry. You may want to
try working on the trees for both you and your match to find the most
likely sources of connection. If you have cousins that have tested, that
can really help narrow things down.

Sometimes the DNA test results can lead you directly to the paper trail
that shows you the breakthrough and the connection. Other times you do
the research as far as you can, but then have to wait until you
eventually find another clue that makes the connection. And sometimes
you run into an NPE which can be really difficult to untangle.

Ancestry is remarkably good at taking care of false positives for
Jewish DNA results, at least through the upper parts of the 3rd-4th
cousin matches. The area that I still wonder about though are matches
with 100 cM or less but many more segments, say 10 or 12. Those seem
to me more likely to be over-estimated at Ancestry.

Best regards,

Michael Good
Los Altos, California


Dorann Cafaro
 

I agree with Raina that you most likely have a common ancestor. I would
also suggest that you take a closer look at your branches further up the
tree to see if you can find a common location. You may have a sister
without a maiden name or a changed surname at the root showing
completely different surnames on their branches. Enjoy the hunt.

Dorann Cafaro

From: Raina Accardi <raina.accardi@outlook.com>
Date: Sun, 5 Aug 2018 12:05:16 +0000

Hi David,

Its hard to tell >from the Ancestry relationship estimate exactly what is
going on but an extremely high probability and that much shared DNA is
not a false positive. You two are definitely cousins of some kind. If
you can convince the match to upload to Gedmatch you will be able to get
a better estimate of the relationship by looking at the size of the
individual segments (larger sizes will indicate closer relationships) and
see if there are a lot of small segments (indicates endogamy). A 4th
cousin would share a 3x great grandparent so you may be looking at
roughly 150 years ago in the 1860s. Can you find any common locations
between your matches ancestors and yours at that time? If so, you might
be able to infer who the common ancestors are even if their surnames are
different. Many Jews took different surnames even among the same family,
some after the mother, some after the father, some made up new ones.
Another course is to look at the family trees of "shared" matches. Are
there any names and locations that connect to either your or your
match's tree? If so, that may be a line of further research.


rayvenna@...
 

Another thing to keep in mind is that many people changed their
surnames. My tree only goes back to the 1830/1860s, yet I've
documented at least three surnames used by one of the male lines in
additional to the usually variations in spellings. One man took his
wife's surname. Another had some trouble with the law, so he changed
his name to hide his "shame." Sometimes people just adopted the
surnames of the cousins who sponsored them when they immigrated, etc.

A common weak spot in most of our trees are maiden names, particularly
of those >from early ancestors. A DNA test doesn't distinguish between
someone who is a 4th cousin through a female line where the surnames
are less likely to match and one along a known male line.

I would definitely consider transferring your test to Gedmatch or to
FTDNA where you can look at segment lengths. Ancestry is great because
of its huge database, but it doesn't have the best tools for looking at
endogamous populations.

Mindie Kaplan

----- Original Message -----
From: "David Brostoff davbro@earthlink.net"
Sent: Sunday, August 5, 2018 1:13:20 PM

On Aug 4, 2018, at 7:09 PM, David E Goldman lugman@verizon.net wrote:

My known last names go back to my great-great-grandparents, and
even a couple more beyond that.
Have you and your match identified all your great-great-aunts and their
married names?

This is usually the limiting factor.