Sunday, February 11, 2018

Ashkenazic Jewish Shared DNA Survey - Initial Results

Updated data here.

Thank you to everyone who has contributed data about shared DNA in people with Ashkenazic ancestry!  I have 4000 data points to analyze, and that should help the entire genetic genealogy community  (I'm still collecting data--you can find out more about the project and how to contribute here).

So far I've been going through and finding anomalies in the data and contacting those submitters (if they left email addresses) to clarify some things.

I'm going to do some deeper statistical analysis once I get the data cleaned up, but meanwhile, here is a high-level look at the data so far.  This data contains entries regardless of the percentage of Ashkenazi DNA (later posts will break down shared DNA based on how much Ashkenazi ancestry the testers have).  This is using very rough data, but I wanted to get some initial information out there to hopefully help inform people while I work on cleaning the data and then analyzing results.

RelationshipNumber of submissionsMin total shared cMMax total shared cMAverage total shared cMMin largest segmentMax largest segmentAverage largest segment
Parent/Child154337937303524.72203.4284.26269.23
Siblings170224630372677.7756281.5185.91
Half Siblings29128621671819.2479.4253.7150.87
Grandparent/Grandchild25136823121817.57114.5263.7187.27
Great Grandparent/Great Grandchild76181001785.566712697.93
Great-great Grandparent/Great-great Grandchild1511511511888888
Uncle or Aunt / Niece or Nephew20167123031786.0335268.8125.14
Half Uncle or Half Aunt / Half Niece or Half Nephew17763.31171936.3984.5161.5114.65
Great-Aunt or Great-Uncle / Great-Niece or Great-Nephew36363.32034.6863.127237.693.08
Great-Great-Aunt or Great-Great-Uncle / Great-Great-Niece or Great-Great-Nephew6620.31019.5855.5252.110774.83
First cousins3103831475.4923.741822183.7
First cousins once removed43361.21878493.7215133.259.7
First cousins twice removed510537263.45079.142.4
First cousins three times removed1072170.1110.61175226.94
Half first cousins17192841.2511.5224150.770.48
Half first cousins once removed2963.7389.6264.0823114.652.94
Half first cousins twice removed582.2212156.843675.559.97
Half first cousins three times removed234.754.844.759.210.49.8
Second cousins70442.4698275.19642347.65
Second cousins once removed6610446169.850565.137.72
Second cousins twice removed1040285.7113.91078.529.05
Second cousins three times removed18017799.9203220
Half second cousins6830313153.911.885.537.94
Half second cousins once removed85024387.61060.523.84
Half second cousins twice removed13091.840.43035.116.22
Half second cousins three times removed230.547.939.212.322.817.55
Third cousins3460315116.9109126.6
Third cousins once removed3200211.27806820.16
Third cousins twice removed43017365.9040.815.67
Third cousins three times removed1464646000
Half third cousins700101.930.04032.19.37
Half third cousins once removed21079.317.65018.85.07
Half third cousins twice removed1108108108101010
Fourth cousins219066449.4055.613.44
Fourth cousins once removed104014633.76041.510.14
Fourth cousins twice removed60102.458.57031.210.87
Half fourth cousins30080.124.36067.914.22
Half fourth cousins once removed6062.133.42062.121.48
Half fourth cousins twice removed1109109109666
Fifth cousins30011628.25022.54.67
Fifth cousins once removed260129.539.52028.18.19
Sixth cousins50109.547.8202811.16
Sixth cousins once removed60113.876.4703217.07
Half sixth cousins27511494.5555

Note that some relationships don't have enough data to really do anything statistically.  But as more people enter data, perhaps they will reach that threshold.

Remember, you can still contribute information about your own family's shared DNA to the project. Information on this project and how to participate is here.

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

15 comments:

  1. This is really interesting, Lara. I haven't checked, but I assume you have---how different are these numbers from the average expected shared cMs/largest segments of non-Ashenazi test takers? Do these numbers reflect what we'd expect from an endogamous population?

    I hope you get more people to submit. Looking at my own known relatives (whose results I submitted), they all fall near your averages. Thanks for doing this!

    ReplyDelete
    Replies
    1. The numbers are definitely noticeably higher than the Shared Centimorgan Project for all but the closest relationships. Stay tuned!

      Delete
  2. Hey Lara, clicked too fast, sorry. Sent you 29 tonight after about 30 yesterday but please delete the first one I sent tonight -- 1/2 sibs, 1810/176/43 100%/100%.

    Gil

    ReplyDelete
  3. Replies
    1. Not that I'm aware of. However, as I mentioned in my initial posting about this project, it's something I'd like to do in the future. First I need to deal with this batch of data though!

      Delete
  4. How can I print this out so it is large enough to see-when I minimize the chart, then screen shot, it is too small. Thanks.

    ReplyDelete
    Replies
    1. Sorry, it depends on the specifics of your printer, so I don't know.

      Delete
  5. Several things:

    1. You can paste this into a google spreadsheet and then sort and recompute to your heart's content. My guess is that you can do this in Excel, too.

    2. I'd have loved to have seen double cousins, etc. Quite possibly, separating double first cousins from of "ordinary" first cousins would bring the "ordinary" cousins' results into closer agreement with the general population. I have quite a few double first cousins in my tree, FWIW. (Since you have halves, why not doubles? After all, it's part of what endogamy is all about.)

    3. I'd have loved to have seen min, max and avg number of segments. I realize that this isn't as good a matching criterion as maximum segment length, but it could still be useful to eyeball matches on ancestry. For a constant # of cM, I'd expect this to go down as maximum segment size increases.

    ReplyDelete
    Replies
    1. More stats to come. I just need to find the time.

      Delete
  6. How do you count the number of segments? I am using the chromosome browser in Family Tree DNA. If a segment appears to cross between the left and right hand sides of the chromosome, does that count as one segment or two?

    ReplyDelete
    Replies
    1. On the upper left where it lists who you are comparing with, it says the number of segments.

      Delete
  7. Hi Lara, I am curious as to whether data is still being actively collected and results updated. Thanks in advance.

    ReplyDelete
    Replies
    1. Yes, I am still collecting data, and I did additional posts after this one (see link at the top of the post). When I have enough new entries (and enough spare time), there will be yet another update.

      Delete
    2. Many thanks, Lara. I was aware there was a 2019 update, but wondered about additional progress. This is invaluable work; I hope to encourage members of my local JGS to contribute.

      Delete