Monday, August 21, 2017

A Relationship Outlier

Blaine Bettinger runs the Shared Centimorgan Project in which he's collected thousands of data points for amount of DNA shared between people with known relationships.  He can give a range of DNA which has been observed to be shared for people with specific known relationships.  If you have a new DNA match, you can see how much DNA you share with that person and use Blaine's chart to identify potential relationships that are possible with that person.
Shared cM Project.  From  Used with permission.

Until now, all of my tested cousins have fallen within the ranges Blaine gives (generally on the higher end due to endogamy)--until now.

When I was in Michigan over July 4th visiting the cousins we originally learned about via DNA, I brought along some DNA kits to get some more data on Diamond DNA.  And the results are starting to come in.  One person I tested was my cousin Dave (father to the Dave who I initially matched back in 2013--who was the key to reconnecting with this branch of my family).  This Dave is a second cousin to my father, so based on Blaine's chart, he should share somewhere between 43-504cM with my father.
Dave's FTDNA match to my father--631cM of shared DNA
Dave wasn't in that range at all.  FTDNA said that he shared 631cM with my father.  I uploaded Dave's kit to GedMatch and looked at how much DNA Dave shared with my father (using a threshold of 7cM):
DNA shared by Dave & my father
GedMatch has the pair sharing even more DNA--642.8cM, including huge segments of 93.9, 66.2, 62.1, 35.2, 32.3, & 29.1cM.

Now, Dave is only 50% Ashkenazi, so there is likely no DNA from his father contributing to shared DNA with my 100% Ashkenazi father.  But how about Dave's sister Beth?  How does she match my father?
DNA shared by Beth & my father
Beth shares only 313.7cM with my father--less than half of what her brother Dave does and well within Blaine's range for second cousins.  (And yes, DNA shows that Dave & Beth are full siblings.)  It seems like through the randomness of DNA recombination, Dave & my father simply happened to get a lot of the same Diamond DNA.

I put together a McGuire Method diagram (a great way to visualize DNA that Lauren McGuire invented--you an read more about this method here) for tested descendants of my great-great grandparents Hillel and Hinda Diamond.  You can see how much DNA each tested person in this chart shares with all his or her tested relatives.
McGuire Method Diagram - check out Dave's match to my father in the yellow box!

It's quite interesting how much more DNA Dave and my father share than do Dave and my father's siblings/first cousins or my father and Dave's sister Beth.  What's also intriguing is that Uncle Leibish's grandson Dan shares a lot of DNA with descendants of my great grandfather (my father, his siblings and his first cousins) but nowhere near as much with descendants of my great grandfather's sister Jenny.

Look at the variations of amounts of shared DNA between second cousins in this chart alone.  This shows why testing multiple people--even siblings--can be very informative.  You never know which will be your Dave.

Note:  I'm on Twitter.  Follow me (@larasgenealogy).

Want to get future blog posts emailed to you automatically?
Enter your email address:


  1. Very Interesting! I see what you're saying about the randomness of DNA recombination. My great-aunt had a match with a woman with 86 total CM, longest segment being 40 CM's. FTDNA predicted their relationship as 3-5 cousins so this seems to be in line. However, my dad and I (1-2 generations below) barely match this woman at all!

    1. Forgot to mention my great-aunt and the woman are 4th cousins (by paper-trail)

    2. Yup--some may get lots more than expected, but others may get much less!

  2. Thank you for sharing this. I have more cousins to test (I purchased half a dozen tests during the most recent sale at several testing sites) This gives me even more reason to get those cousins tested.

  3. I wonder if anyone has created a "Shared DNA" table that also looks at longest segment? In my experience this seems a better predictor than total amount of shared DNA.

  4. You say "DNA shows that Dave & Beth are full siblings", but your chart shows 1687 cM shared which is too low for full siblings and consistent with half-siblings

    1. That's a typo (I just double-checked!) and should be 2687cM. I'll update it tonight. Thanks for catching it!

  5. A friend submitted a first cousin match who was half what the average should be. Suspect we will see a lot more outliers than the numbers suggest with all the testing going on and as more grandparent - grandchildren and first - third cousins test.

    1. The first question is whether your friend is sure that the person isn't actually a half first cousin. But it definitely could be an example that's the opposite extreme of my father & Dave!