Re: DNA Matches #dna

David Ellis

David Goldman ( writes:

I hate to say that I think the matches I am provided through the
various sites are driving me batty. I realize that in most of them called
"third cousin" the longest segment is only around 10 at the most, with
the "overlap" varying below and above 100, but why does the system proclaim
a cousinhood of third cousin or such when this surely doesn't mean that
in most cases?!! All it means is that two people have some common genetic
background at some time in the distant past that is retained in us,
even where some of my matches are with non-Jews who probably had had a
Jewish ancestor in the distant past. But this doesn't translate into
a third or fourth cousin.

Welcome to the wonders of Ashkenazi Jewish endogamy. We are all descended
from a founder population of about 350 people some 700 years ago, when our
numbers hit a bottleneck after the Crusades, the Black Plague and pogroms.
Staying within our own group, we married our first, second or third cousins
for many generations until our population spread and grew much larger in the
18th century. As a result, unlike other European ethnic groups, we share
many common DNA segments >from centuries ago. The DNA experts don't have a
great handle on how to distinguish these >from the segments that legitimately
correspond to genealogically reachable common ancestors.

The vast majority of the DNA matches estimated as third or fourth cousins
are actually quite a bit more distantly related than the results imply.
Hidden among them are a small fraction of sufficiently recent and accessible
relatives. How does one find the needles in the haystack? A rule of thumb
I use is to follow through with matches that share over 100 cM, with a long
segment at least 20 cM and another segment at least 10 cM. It's not
foolproof, but it helps identify promising leads.

Genetic genealogy is still very much a developing field, and the situation
can only improve over time.

David J Ellis
Natick, MA 01760

Join to automatically receive all group messages.