Sunday, January 14, 2018

Ashkenazic Jewish Shared DNA Survey

Ashkenazic Jews and those with Ashkenazic ancestry often share significantly more DNA with one another than would be expected because of endogamy; Ashkenazim all descend from a small group of people whose descendants all married one another--so we're all cousins many times over.  I've looked at how endogamy manifests in my own family (here, for example), but there are many others who have tested pairs of known relatives who could contribute data to our understanding of how having Ashkenazic DNA impacts amounts of shared DNA.

[There will likely be follow-on surveys to gather similar data on Sephardic and other Jewish subgroups.]

I was inspired by Blaine Bettinger's Shared cM Project which looks at distributions of shared DNA between people with specific known relationships.  But Ashkenazim don't always fit into Blaine's bins because of endogamy.  Getting a better understanding of how having Ashkenazic DNA impacts shared DNA should help those who are trying to understand how close an unknown match actually may be.

So I need your help to let me know how much DNA your known relatives who have DNA tested share with one another.  The more sets of known relatives and information we have about the DNA that they share we know, the better we will be able to understand how endogamy impacts how much DNA Ashkenazim share with one another.

Please fill out this form with data for each pair of known relatives of yours who have tested.  You'll need to fill out the form separately for each pair.  Please only fill out the form if both individuals are related through shared Ashkenazic ancestry.  Aggregated information collected from this form will appear in a future blog post, so stay tuned!

Where to find the requested information is as follows:

At Family Tree DNA, for example, you can find the numbers here:


At AncestryDNA, you can find the numbers here:
At GedMatch.com, you can find the numbers here:
At MyHeritage, you can find the numbers here:


For 23AndMe, first go here and select the relative for comparison and click "compare."
Then scroll down to the bottom to find the segments; choose the largest one.


You can contribute HERE!

This project will benefit all genetic genealogists with Ashkenazic Jewish ancestry; the aggregated information (minus emails) will be analyzed and will be posted on this blog.

Thank you for participating!

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


61 comments:

  1. You may want to fix the hole between 1/2 (2 grandparents) and 100%. For example, where would you answer for 3 grandparents, or for all but one great-grandparent?

    ReplyDelete
  2. Lara - I made an error on one of the entries I have just put in, but I can't see a way to go back and correct it ...

    ReplyDelete
    Replies
    1. Email identifying info (maybe what the total & max were that you entered), and I'll delete. Then you can re-enter the correct info.

      Delete
  3. I have tested against the same 3rd cousin on ancestry & 23&me. On ancestry we share 49cM across 7 segments while on 23&me we share 82 cM across 7 segments. I had never noticed this before. This will make comparisons for your chart difficult to work I think.

    ReplyDelete
    Replies
    1. That is why I ask which service the comparisons are done with. Hopefully I'll have enough data to break things out by platform.

      Delete
  4. FTDNA has tested (atDNA) 8 of my family members, including myself. The closest relative is an uncle. Mostly 1st & 2nd+ cousins. To do what you ask appears to be 25 pages to fill out!! (7+6+5+4+3+2+1) I'd like to be part of your study. Is there any way to cut his number down a bit??

    ReplyDelete
    Replies
    1. You can do a subset of those if you want. Or maybe do 5 a day, and you'll be done by the end of the week? I hope you'll participate!

      Delete
  5. Did my cousin last after doing my 5 kits I manage and forgot to put in my email. It had 724 shared cM and 31 segments shared.

    ReplyDelete
  6. I've entered the data for my various kits. I'm a fan. Best Wishes!

    ReplyDelete
  7. How about a 3/4 2nd cousin? We are 1/2 2nd cousins (share ggf) and 1/2 3rd cous.ins (ggms were sisters)

    ReplyDelete
    Replies
    1. Yes, as the instructions on the form say, choose the closer relationship (1/2 2C) and then choose “yes” for multiple relationships and explain that the individuals are both 1/2 2C and 1/2 3C.

      Delete
  8. Trying to understand this....I have tested 8 relatives (sibling through cousins once/twice removed). The only easy pair is my daughter and her father. Am to compare each of my tests against myself, then each of them against all the others? I can manage it, just don't want to do more work than necessary!

    ReplyDelete
    Replies
    1. Yes, if you can compare all of those against one another (assuming they're related to one another as well), that would be ideal and give us more data.

      Delete
  9. Number of Shared Segments: In the Chromosome Browser FTDNA shows shared segments with 1+ or 3+ or 5+ or 10+ cM. I assume we use the default, 5+ cM?
    (do they all use these numbers the same way? it not, how useful is it?)

    ReplyDelete
    Replies
    1. Please use the default regardless of platform. Assuming I have enough submissions to make it reasonable, I'd like to break out by platform because yes, there are differences.

      Delete
    2. Most of my cousins tested at FTDNA where I believe the default is 5. I and my most distant known cousin originally tested on Ancestry which is the only testing site he is on. Since I could not get the largest segment there I also took a look at how we compared on Gedmatch where the default is 7cm. That yielded a smaller total cm than Ancestry (145.4 vs. 151) in 8 segments vs 9 on Ancestry, but when I dropped the Gedmatch threshold to 5 (as all my other results will be showing) the total cm jumped to 162.4 and segments to 11.

      Should I report this match twice, once using Ancestry without largest segment and again from Gedmatch using the 7cm default. Is the 5cm result useful in some way when comparing to FTDNA matches?

      Delete
    3. Pick one and report that. Preferably GedMatch over Ancestry since you get more data (such as largest segment) there. Thanks!

      Delete
  10. Entered my 20 pairs of "cousins". I think I said "yes" about some of the question on multiple relationships. If there is no explanation then it should be "no"
    Also, all of them say 100% Jewish. I will go over them and let you know if any should be changed.

    ReplyDelete
  11. Lara, I think that I made a mistake on one of your survey questions, it is the one that said 5th cousins. Could you please tell me the shared CM and the Longest block? I think it was a third cousin relationship. My email is sarahlmeyer@suddenlink.net.

    ReplyDelete
    Replies
    1. The one that says 5th cousins is 94/10/24. So you're saying it should be 3rd cousins?

      Thank you for doing so many entries!

      Delete
  12. Lara,
    I have finished my FTDNA, but I would like to check my responses. Could you send me a copy of the google docs spreadsheet with just my responses. I need to double check some of the known relationships. I am sure that the numbers are right. I was very impressed by your form, but after filling it out somewhere between 60 and 80 times, it would have been helpful to have a column for comments useful to the submitter - where I might have written sarah and susan so that I could more easily check things.

    ReplyDelete
  13. Lara
    Just sent @ 30 entries. Looking forward to your findings. Good luck

    ReplyDelete
  14. Thank you for undertaking this project. Just the task of organizing my information has taught me a couple of things.
    I have tested my mother and her two brothers and several cousins on FTDNA. I was surprised to see that my mother is not genetically matched to a known 2nd cousin once removed. Both her brothers are matched to him. I think this is useful to your project, but I'm not sure how to enter this zero amount information. What would you like me to do?

    ReplyDelete
    Replies
    1. You should be able to enter 0 in those fields. Thanks for participating!

      Delete
  15. Hi Lara. I have had 14-15 relatives of mine tested. Many of them are related to each other, as well as to me. I can't do the permutations, but if I were to fill your form out for each relative and me, as well as each relative and every one of their relatives, it would take a whole lot of time. I'm happy to give you my Gedmatch login and a who's who guide but I don't have time to do all the data entry. madeleine.sann@gmail.com

    ReplyDelete
    Replies
    1. Thanks, but I already entered numbers for the 60-ish people I've tested, and I'll be spending quite a bit of time analyzing the submitted data. If you would like to participate to help, I'd appreciate it; if not, I just don't have the time to do data entry for others.

      Delete
  16. Hi Lara,
    I've got a similar situation as Madeleine. I've tested 30+ relatives through 4th and have created a spreadsheet with their info I can send. Plus, I have created a "Minimum Criteria Guide" I ask people to use to help them determine how close a real connection is. It's been pretty accurate through 3rd Cousin if you want to pass on to others. It actually might be the answer you are looking for.
    Best,
    Elyse

    ReplyDelete
    Replies
    1. Elyse,

      I tried a spreadsheet with a few people, but I ended up having to manually fix things because of formatting/nonstandardization, and I simply don't have the cycles to do this for everyone's spreadsheet.

      Anything you can input into the form would be appreciated.

      I have a criteria I use (see https://larasgenealogy.blogspot.com/2017/03/a-technique-for-endogamous-dna-using.html), but this project is to see how likely you are to have outliers from that.

      Delete
    2. I took a look at your link. I've narrowed down the criteria quite a bit from that using my own matches up to 4th cousins on maternal & paternal sides. Here's my guide which has helped me find some actual matches and saved others a lot of time. I first sort normally, then by largest cM block.

      Minimum Criteria using GedMatch: AT LEAST one shared cM block longer than 35; PLUS a minimum two to three blocks at least 10-15 cMs; PLUS total cMs around 100; PLUS a GM estimate no more than 3.5.

      Delete
    3. You will miss some decent matches that way. This study shows that (especially with that 35).

      Delete
  17. I have too many relatives and not enough time to fill out the form but am happy to post my Minimum Criteria Guide letter that I send if it might be helpful. Let me know.

    ReplyDelete
  18. I would like to participate but I cannot find the information you ask for on 23 and me. Can you help?

    ReplyDelete
    Replies
    1. Look at the very bottom of the post, and there are instructions--a link to click and then where to go/look.

      Delete
  19. how can I correct a submission, just remembered multiple relationship

    ReplyDelete
    Replies
    1. Please comment here with info that can help me identify which submission as well as what should be corrected.

      Delete
  20. I'd like to contribute but I am one of those trying to evaluate relationships from an unknown ancestor using shared centimorgans! I think I have identified who it is and what our relationship is to the Ancestry generated 2-3 cousins, however I wouldn't want to add that data until I have the confidence of all those involved. I wonder if you would consider looking at my research and giving me your opinion?

    ReplyDelete
  21. If I'm using the data from GEDmatch, do I enter the numbers from the one-to-many comparison chart, or from the one-to-one comparison? Your image shows the one-to-many, but the one-to-one is the only place to find the number of segments, so I'm not sure if using the numbers from the two different comparisons would mess things up.

    ReplyDelete
  22. I'm not sure that you would really want my info. for your study. On Gedmatch I have about 200 matches on ch.8 which are Ashkenazi as far as I can tell. I haven't yet analyzed my other chromosomes yet because I am still working on this list and triangulating. These matches would be distant. My paternal grandparents were born in Mexico. My mother had one grandmother born in Germany. The rest of her ancestry seems to be Eng/Irish. What do you think? Thank you for your time.

    ReplyDelete
    Replies
    1. For this, I want people who know their percentage of Ashkenazi. So while your data won’t help, hopefully you will benefit from the results.

      Delete
  23. On my heritage I can see percentage of Askenazi for my matches. Do you want that information or only my known tested relatives? My mother has none but my children each have some.

    ReplyDelete
    Replies
    1. Only known tested relatives with documented Ashkenazic ancestry please.

      Delete
  24. You've probably already thought about this, but here goes anyway: it could be valuable to researchers to conduct a similar sort of data gathering survey on the amount of dna sharing between known double cousins within the Ashkenazic community.

    ReplyDelete
  25. Trying to figure out a bit of a mystery here and looking for some help! My dad did a test through 23&Me and matched with a woman who shares 16.2% DNA, with 1207 cM & 26 shared segments. The longest of the shared segments is 226.20 cM which I know is extremely high. The woman he matched with does not know who her biological father is. My dad is 98.6% Ashkenazi Jewish and she is 47.6% Ashkenazi Jewish.

    Trying to figure out if her father is/was a full or half brother to my dad whom we just don't know of for whatever reason or if it's possible that my dad's dad had a child with one of his sisters and that male child was given up for adoption or something?

    One of my dad's first cousins has done her DNA through Ancestry and I'm wondering if I should just ask her to share the raw DNA file with me so I can compare it as well.

    The woman he matched with at 16.2% unfortunately cannot ask her mom about any of the history as she has dementia. But the relationship is obviously through her paternal line.

    We're just trying to figure out who her dad is. Obviously it's a close relation due to the long strands of shared DNA, we would just really love to understand more! She didn't even know she was half Jewish until she got her results back, nor did her or her father that raised her know she was not his child. Crazy stuff and slightly confusing since we don't know who her father is, if he is still alive and how to even go about trying to find out any of those answers.

    I'd love to get any feedback from you! Thanks!

    Katherine
    nelehkat@gmail.com

    ReplyDelete
    Replies
    1. It's more likely that the person and your dad are half aunt/uncle and half niece/nephew. But half siblings or full aunt/uncle aren't out of the running.

      You should definitely see if that first cousin will upload her DNA to GedMatch Genesis, and you and the mystery woman should do the same, so you can all compare as well.

      Delete
  26. Lara, I previously submitted some entries for your tables. At present, 35 of my known relatives have been tested. Is there any way you can tell me who I've entered already so I can submit the rest to you?

    ReplyDelete
    Replies
    1. Unfortunately I don't capture names as part of the survey.

      Delete
  27. I entered info for my 3 children, full siblings, who all have 1 Ashkenazi grandparent. But after submitting, I wonder if I was supposed to add up Half Identical and Full Identical CMs and segments? Because I only put in Half Identical. I couldn't find any instructions on this.

    ReplyDelete
    Replies
    1. Just whatever number is given as the overall number. Thanks for contributing!

      Delete
  28. I must have overlooked this. You have my info and my/our Supkoff relatives. Do I need to repeat? I think I have two from my other side.

    ReplyDelete
    Replies
    1. I've done the Supkoffs, but if you have from the other side, please fill out the form with their info! Thanks.

      Delete
  29. Lara, am delighted to find this. (However, if I've done it already & completely forgotten, let me know, please. I do that sort of thing.)
    How do you avoid the relationship between 2 individuals being entered twice, ie by each individual?

    ReplyDelete
    Replies
    1. I don't know if you did, or didn't do it already....

      Since I don't ask for identifications of the pairs of relatives, it's possible that the same pair could be entered twice by the two individuals, but I think it's prety unlikely.

      Delete
  30. Have you published a chart that we can look at? Also, have you included people who have discovered partial Ashkenazi ancestry? Thanks!

    ReplyDelete