Ashkenazi DNA: we are all descended from 350 people in medieval times

Most of my DNA explorations have centered around my Dad’s Norwegian DNA because so many Norwegians have tested and the populations of those ancestors are only mildly endogamous; so it is easy to find new cousins and fun to work with those results.

My grandfather and me 1955

My grandfather and me, summer 1955

My mother’s father was a German Jew. The number of DNA matches this gives me, my brother, and my two maternal aunts is astronomical. Frequently I will match someone from Eastern Europe for five or six large segments who cannot share an ancestor with us for the last 200-300 years and is even listed as a “distant cousin.”  If they were Norwegian, that amount of shared DNA would make them my 2nd cousin.

This has been so frustrating that I just about stopped working on my Jewish DNA. A fellow Jewish researcher told me to ignore matches who did not have at least one 23cM shared segment!

Today there were dozens of news articles about the European Jewish founder effect suggesting that all Ashkenazim are descended from about 350 people who lived in about 1300 A.D. or so. This, combined with a fair amount of endogamy, would explain the large amount of shared DNA among European Jews.

These articles are all based on a study published online at Nature Communications ( by Shai Carmi with a group of people including his mentor, Itsik Pe’er, a cofounder of the Ashkenazi Genome Consortium (TAGC – notice those letters!). Professor Pe’er is an associate professor of computer science at Columbia and a consultant for 23andme.

TAGC has collected the full genomic sequence for 128 healthy ashkenazim from many cooperating sources to create a data resource for studying diseases specific to this group. However this collection has far more uses, including the study of the founding population of European Jews.

Here are some of the key findings and points of that published study:

  • The founding population for Ashkenazim of a few hundred in about 1300 A.D. was about half middle eastern and half European.
  • Europe was largely repopulated by people from the near east after the last glacial maximum.
  • There are more disease causing mutations in Ashkenazi DNA than in European DNA, an effect of the small founding population followed by genetic isolation.
  • This data resource, the 128 genomes, will allow better understanding of jewish mutations that are not harmful as well as the disease causing ones.

The article that best summarizes this study for us laymen is this one from the Times of Israel:  Enjoy!


6 thoughts on “Ashkenazi DNA: we are all descended from 350 people in medieval times

Click here to add your thoughts at the end of the comments
  1. Kitty, so even if there are multiple chromosome matches, if there are not any greater than 23cm you ignore it? That is good to know, as we are finding it difficult to find many matches that long. Just like the comparison my husband and you have on GedMatch.

    Thanks for a great article.

    • Suzette – you got it! 5 or 6 segments > 6cM and less than < 12 cM are pretty meaningless in my experience. But I pay attention when there are several segments > 10cM although I have yet to find a confirmed match on my Ashkenazi side. I have one set of folk with ancestors from my ancestral town of Floss, Bavaria that I am content are true cousins of some sort. My known double 3rd cousin shares segments of 30cM, 22cM, 13cM, and 9cM plus many much smaller ones. See that in this blog post

      So I am working on a blog post now with some help on how to proceed with Ashkenazi DNA. My jewish grandfather was an only child so there are no 2nd cousins for me to work with. A number of known 5th cousins have tested however as well as my double third.

      • I’ve been meaning to finish my blog entry about Polynesian endogamy to help others understand how to determine a true 1st – 2nd cousin match.

        Just last week Friday the DNA results of my mom’s half 1st cousin were ready from FTDNA. Just as my cousin took this DNA test & sent it in is when my mom admitted that she remembers as a child being told by her mom that she was adopted. I only had this cousin test to verify if this was true or not.

        On FTDNA my mom & this cousin shares 501.35cM, predicted 1st – 2nd cousin. A half 1st cousin would share about 425cM. On GedMatch, they show 216cM, 24 segments. The largest segment is only 14cM (GedMatch), but I’ve learned that although my mom can have matches above 200cM, the largest segment averages around 15cM. On 23andme, I can tell my matches although many 2nd & 3rd cousins, they have around 15+ segments, definitely not a true 2nd – 3rd cousin relationship.

        So I am now teaching people about recognizing the number of segments and segment size over the total as a better prediction of gauging how real of relationship it is. Plus I have examples of known 2C1R, 3C and 1C2R in our endogamous group to compare to.

        • Kalani that sounds fascinating. Please finish your blog post and share it here – you can email me, I am kitty at this blog’s url without the blog. in it

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.