Second Cousins (Or Closer) That Don’t Share DNA?

Caution should always be taken before pursuing an answer to a genealogical question. Ask yourself (or your relative) before diving into the research whether you really want to know the answer, and whether you think you are prepared for it. Just because you as a genealogist may be prepared for the result does not mean everyone in your family will be.

This is as true with DNA evidence as it is with any other type of genealogical evidence. Just as a census record or birth certificate can reveal a family secret, so can the results of a DNA test. The Genetic Genealogy Standards, for example, make clear that unexpected results can occur:

12. Unexpected Results. Genealogists understand that DNA test results, like traditional genealogical records, can reveal unexpected information about the tester and his or her immediate family, ancestors, and/or descendants. For example, both DNA test results and traditional genealogical records can reveal misattributed parentage, adoption, health information, previously unknown family members, and errors in well-researched family trees, among other unexpected outcomes.

This is especially true when testing two people who are second cousins or closer, as there has NEVER been a single demonstrated case of second cousins or closer who fail to share DNA. Let me re-phrase that: no one has ever demonstrated second cousins or closer that fail to share DNA at a level detectable by one of the three testing companies. There are claims out there, but they are completely unsupported at the current time. (Also see “Are There Any Absolutes in Genetic Genealogy?” for more).

What would be needed to prove a case of actual second cousins or closer that don’t share DNA? You would need ALL of the following elements; without any one element, you have NOT proven the case:

  1. Test results for second cousins or closer that don’t share DNA according to 23andMe, AncestryDNA, FTDNA, and/or GEDmatch (if at AncestryDNA, the result would need to be confirmed at GEDmatch to ensure that Timber isn’t having an impact); AND
  2. Test results for each of the related parents of the second cousins or closer, showing that: (1) the two parents share detectable amounts of DNA in common with each other at a level commensurate with their genealogical relationship; (2) that parent 1 matches child 1 at the parent/child level; AND (3) that parent 2 matches child 2 at the parent/child level.

So, in the following diagram, if first cousins Angela and Bill don’t share DNA, then the following elements would be necessary to support the claim that they are actually first cousins rather than a case of misattributed parentage:

  1. Angela and Bill do not share detectable DNA;
  2. Abel and Bruce share DNA at an appropriate sibling amount;
  3. Angela and Abel share DNA at an appropriate parent/child amount; AND
  4. Bill and Bruce share DNA at an appropriate parent/child amount.

picture1

Without any of the above elements, you have NOT proven the case. This is a very high burden, but that is required to establish that the case isn’t actually the MUCH more common occurrence of misattributed parentage. After all, Occam’s Razor suggests that misattributed parentage (no matter how much the parties may protest) is significantly more likely.

If someone tells you that they have a case of second cousins or closer not sharing DNA, ask them for the GEDmatch kit numbers, and confirm that the above requirements are met. Otherwise, be a skeptic. Unfortunately it seems sometimes that our culture has taught people that skepticism is unhealthy, so fight that urge. Be a skeptic, ask for the evidence!

 

 

P.S. – note that it is technically possible that a second cousin or closer wouldn’t share DNA, but the odds against it are astronomical. You likely have a significantly better chance of being struck by lightning twice in your lifetime (1 in 9 million).

33 Responses

  1. Lydia Bishop 3 October 2016 / 10:01 am

    Very interesting. I have a situation similar to what you discuss here. I have a known paper trail second cousin who comes in at the FOURTH cousin level. My known paper trail fourth cousins and I share more DNA than my second cousin shares with me. All these cousins are on AncestryDNA.

    My grandfather and “DF’s” grandfather were brothers. Other second cousins, descended from my grandfather’s other siblings are within the expected range. “DF” and some of the other second cousins appear to be in the expected range. Just not me and “DF.”

    Unfortunately these cousins don’t want to transfer to GEDMatch for further study. I am sure that if “DF” were to transfer to GEDMatch his shared DNA with me would be in expected range. We’re a “victim” of Tinder’s algorithms.

    Guess I’m stuck.

    • Andy Baker 3 October 2016 / 1:26 pm

      Lydia, I am positive you mean Timber (Ancestry’s matching algorithm), but it would certainly be interesting if cousins were matched on Tinder (the dating app). 😉

    • James Owston 3 October 2016 / 4:14 pm

      Lydia: Sometimes two second cousins have a lower than normal match, but they match others in a typical range. I’ve seen this in my project twice with a second cousin pair (82cM) and a second cousin, once removed pair (31.4cM).

    • cathy 2 September 2017 / 6:43 pm

      where can I find someone to test or give me kit to test my fathers 1st cousin and me.
      all the sites I have contacted say the test don’t test any further than 1st cousins.

  2. Peggy Norris 3 October 2016 / 7:32 pm

    I think that there is a “not” missing in the first sentence of the next to the last paragraph. “If someone tells you that they have a case of second cousins or closer sharing DNA, ask…”

  3. Deb 4 October 2016 / 1:29 pm

    I have a second cousin, on paper, that I share no DNA with. Both of us have tested at Ancestry. Our mothers should be first cousins. She also does not match my mother, or my maternal half aunt. Her grandmother was 10 years older than my grandmother and their brothers.

    The absence of a DNA match and the age difference, suggests to me that her grandmother might be adopted. My mother and I have DNA matches to my grandmothers documented family.

    Unfortunately she doesn’t seem to want to investigate further.

  4. Stephen D Echard Musgrave 4 October 2016 / 2:45 pm

    Thanks for the article Blaine, wishful thinking or ignoring the obvious is one of the big blinders to accurate genealogy. Now in my family the men are total rakes and were always finding someone closely related to us who isn’t related to us according to genealogy.
    I have kept this fact discrete among those family members who are not bothered by such things. However it’s a lot more tricky when you’re contacted by someone who thinks mom was a saint while the old dad spent 20 years at sea say in the merchant Marine.

    The only big surprise I’ve had in my genealogy was discovering that huge amount of Cajuns and Acadians I’m related to. For some reason the big block of X DNA appears to come from my maternal grandfather’s great-grandmother’s name was Lejeune. That people though that pop up this fourth cousins etc. don’t usually share any X DNA but I can’t figure out how someone that far back would generate say 300 4th cousi all with roots in Acadia and St. Landry Louisiana

    • Mike Hebert 16 January 2017 / 8:41 pm

      Steven, my mother’s line come from in the heart of where you’re talking about. Your tree should reveal the same surnames popping up time after time, especially among the Non-Acadian families around the Crowley area. Throw in that mix the most common Acadian surnames in Louisiana (from the same area) and you’ll understand why I’ve given up sifting out my 6,800 4th cousins or closer. No mysteries have popped up anyway that I can tell, so I just enjoy being related to every single person alive in Southern Louisiana.

  5. Lisa 5 October 2016 / 2:24 am

    Interesting post. Does the same apply to Second Cousin, Once Removed?

    I understand the randomness of atDNA inheritance and I’ve seen your Shared cM Project estimates but …

    Over at AncestryDNA I have three siblings tested (yes, they test as siblings to each other) each shows a match to a second cousin, once removed — documented by papers and family-gatherings and DNA according to AncestryDNA. (All four were tested prior to the AncestryDNA chip change.)

    Sibling A to Second Cousin, Once Removed
    showed 17.7 cM across 4 DNA segments
    after the May 2016 change 24.0 cM across 4 segments

    Sibling B to Second Cousin, Once Removed
    showed 30 cM across 5 DNA segments
    after the May 2016 change 35 cM across 5 segments

    Sibling C to Second Cousin, Once Removed
    showed 12.8 cM across 3 DNA segments
    after the May 2016 change 13.1 cM across 3 segments

    An Autosomal Transfer was done for all three siblings to Family Tree DNA where again they all showed as sibings with FTDNA saying they had even more cM in common than Ancestry DNA said.

    Second Cousin, Once Removed tested directly at Family Tree DNA a couple months ago. When the results were in — NONE of the three tests showed up as a match! I can see why Sibling C and Sibling A might not show as a match (possibly no segment at least 7 cM) but I thought Sibling B would show up at the very least. Needless to say very disappointing.

    In general this didn’t make sense to me because the mother of the three siblings was directly tested at both Ancestry and FTDNA and came up as the mother with FTDNA saying they had more cM in common than AncestryDNA said. (No, the mother is not the parent related [2C] to the said 2C1R.) Additionally, from prior experience any Matches to the above siblings who have results at both AncestryDNA and FTDNA (whether they transferred or tested directly at FTDNA) all show more cM in common at FTDNA than at AncestryDNA. So I’ve taken to nicknaming this the Timber Effect. I don’t think this FTDNA no match to the 2C1R is because of transferred or direct test as both show the same trend — FTDNA says more cM in common than AncestryDNA said.

    There is one match at FTDNA that matches the three siblings tests and the Second Cousin, Once Removed test but it is a more distant cousin. That one match is also at AncestryDNA and again FTDNA says more cM is common than at AncestryDNA for the three siblings. Not sure yet on the 2C1R test if that match shows as a match at AncestryDNA too and if FTDNA says there is more in common like the other tests. (I guessing it’s the same as my experience with the matches of the Sibling tests but 2C1R can’t return messages for a few weeks.)

    (And yes over at AncestryDNA all three siblings show matches to various known cousins they have in common with 2C1R though none of those are at FTDNA too yet.)

    I’m hoping to get the Second Cousin, Once Removed to upload to GEDMatch. I’m thinking both the AncestryDNA data and the FTDNA data since they were both direct tests though I’m not sure if it will show us what might have happened — why the difference. I’m thinking at this point this is the only way to learn why.

    The parents, grandparents involved in the shared lineage of the siblings and 2C1R are all deceased and can’t be tested. These are the end-of-liners to test and 2C1R has no siblings of the line in common. As I said other known cousins do show as matches in common over at AncestryDNA. So we’re pretty sure there was no questionable paternity issue anywhere. I guess in the meantime I’ll recheck each companies minimum threshold of what a match is or is not.

    Thanks for the post.

  6. James W. Moore 5 October 2016 / 10:03 pm

    If the odds against 2nd cousins sharing no DNA are 9 million to one, then one would expect almost 1000 such cases among living people are earth. So another way to look at your claim would be to observe that it is near certainty that there are a pair of 2nd cousins alive today who share no DNA.

    • Blaine Bettinger 5 October 2016 / 10:14 pm

      Absolutely, that’s why I always phrase it in terms of no one ever demonstrating a case of second cousins or closer who fail to share DNA, not that it is impossible.

      Part of me hopes no one ever finds and documents such a case, because it is so much cleaner to say that so far, no one has ever found a case. If someone does find it, more people will claim it than already do!

      • Stephen D Echard Musgrave 11 October 2016 / 6:34 pm

        Good Blaine just like UFO’s

      • Barbara Taylor 26 October 2016 / 2:20 pm

        It takes much, much more effort to document people who are outliers – those who don’t fit the “expected” numbers. It’s not impossible – just a huge amount of work to confirm. Many other factors have to be looked at than just the numbers when those happen.

  7. Jim Thornhill 21 October 2016 / 9:17 pm

    Could you recommend a book on DNA testing that goes into some detail as to the actual methodology of DNA testing for Genealogy? Things such as how the sample is prepared, what equipment is used, what size sample is needed, margins for error, ect. I have no desire to start doing DNA testing in my garage, but the scientist in me needs more information than I have seen currently available to be comfortable with discussing/recommending DNA testing.
    Thanks

    • Blaine Bettinger 1 November 2016 / 10:15 am

      I’m not aware of any such book, Jim. The companies mostly did not come up with the testing procedures, as they use SNP chips from companies that provide chips for many different purposes. You could start your research by determining the chips utilized by the companies (see http://isogg.org/wiki/Autosomal_DNA_testing_comparison_chart), and then researching the procedures of the company that manufactures/supplies the chip.

  8. JD4x4 26 October 2016 / 2:00 pm

    VERY reassuring to hear the odds about 2nd Cousin or closer matches. How about stats on 3rd Cousins?
    I’m asking because as a result of my DNA tests, at age 64 I was perplexed for almost a year as to why none of my results matched ANY surnames in my extensive family tree. Like others, I doubted the results enough that I retested at another company, alas, with the same results.

    So, even without GEDMatch confirmation (my matches were reluctant to attempt download and upload) I started looking more closely at my matches up to 2nd & 3rd Cousin and those that we had in common. When I started to recreate some of their trees, lo and behold my DNA results began to show me that I was indeed the offspring of 2 people that appear (so far) to have never married (no paper trail).

    I now trust DNA MUCH more than paper records. As I continue to refine what the DNA is telling me, I will obviously try in earnest to get my matches results onto a site that allows for chromosome to chromosome comparison, but meanwhile it would be helpful to hear the stats on 3rd cousin matches so I can better tell who my key matches are.

    • Barb 28 July 2017 / 1:20 pm

      I’m interested in the answer to this as well.

  9. Todd Ronning 1 November 2016 / 10:02 am

    Can I conclude that your phrase “not sharing DNA” is not the same thing as “matching DNA”? (500SNP/7cM) I recently tested against my paper trail 2nd cousin. We did not match, but dropping the SNP to 100, we then match on one chromosome, 7.7cM. I wonder if that is enough to call it “sharing DNA” but not enough to call “matching DNA.” I’m concluding we are not genetically related as we don’t meet the minimum criteria for a match and second cousins should?

    • Blaine Bettinger 1 November 2016 / 10:18 am

      Yes, for example I state that “no one has ever demonstrated second cousins or closer that fail to share DNA at a level detectable by one of the three testing companies.” It’s all about detectable levels, since each company has a threshold.

      That being said, your first hypothesis in your case, based on the probabilities, should be that you and your second cousin are in fact not genealogically second cousins. Thus, it is VASTLY more likely that you have either identified the wrong person or there has been a misattributed parentage event.

  10. Jeff Epstein 30 November 2016 / 4:25 am

    This has been a most enlightening article for me. I am in just such a situation right now. My 2nd cousin & I have only a total of 19.2 cMs on 2 chromosomes (gedmatch). So something’s not right. My father (1C1R) has even less in common w/her. Our grandmothers were sisters. (Or maybe not- that’s 1 possibility). Or it could be just her father was an NPE.
    I’m having another cousin (1st) test for comparison. Also my 2nd cousin’s 1st cousin as well. Any other suggestions would be most welcome from you.

  11. Erin c Van Zante 3 December 2016 / 3:45 am

    Chimeras. They happen enough to have been studied and documented.

    • Blaine Bettinger 4 December 2016 / 9:57 pm

      Very true, we need to keep chimeras in mind when analyzing any DNA results. But in this instance, chimerism won’t have any effect on the match. A second cousin is a second cousin whether they are a chimera or not.

  12. debbe 13 December 2016 / 1:45 am

    Is the reverse possible–can a close DNA second cousin actually not be related? In other words, is there any evidence that testing companies make significant errors?

    I am currently trying to convince the test administrator of my 2nd closest match on AncestryDNA that we have to be related in spite of our well-researched trees showing no surnames, locations or ethnicities in common. The results are:

    Predicted relationship: 2nd Cousins (Possible range: 2nd – 3rd cousins)
    Confidence: Extremely High
    286 cm across 15 DNA segments

    I have strong DNA matches on all recent lines in my family, so I’m confident my tree matches reality. And I have a good working theory on how we probably are related.

    But she is insisting this has to be a mistake–have there been any reports of such errors?

  13. JCB 13 January 2017 / 2:28 pm

    Every bit as interesting as the examples above, I think, are cases where someone very distantly related according to paper trail, turns out to share much more DNA than expected for such a distant relationship. I have a 1/2 4C, 1R (paper trail) who tested with Ancestry & uploaded to GEDmatch. My mom and I tested with FTDNA & uploaded to GEDmatch. The distant cousin wouldn’t be expected to show up as a match to either of us on either website, but on GEDmatch, using default setting of > 7 cM (for all of the data below), this very distant cousin shares 141.1 total cM, largest segment 31.6 cM with my mom (more like 2C, 1R); & shares 105.6 total cM, largest segment 25.9 cM with me (more like 2C, 2R).
    I have an hypothesis of approximately where in the family this “reinforcement of DNA” might have occurred, and have asked one known 3C & one 4C (these are their relationships to me) so far to test at FTDNA & upload to GEDmatch.
    Results for 4C: 10.8 cM largest segment with my mom; 0 cM with me; 0 cM with my 3C; 20.2 cM largest segment with my 1/2 4C, 1R.
    Results for 3C: 45 cM largest segment with my mom; 30.6 cM largest segment with me; 0 cM with 4C; 15.4 cM largest segment with 1/2 4C, 1R.
    We are also X-matches with my 1/2 4C, 1R, who shares 66 cM largest segment with my mom & 65.9 cM largest segment with me. 3C is an X-match as expected to my mom at 28.8 cM largest segment, to me with 28.9 cM largest segment, to my 1/2 4C, 1R with 32.0 cM largest segment. 4C has no X-match to any of the rest of us, as expected (paper trail). Testing another known 3C and a 2C might help pinpoint the location of the NPE, IF they would agree to test.

  14. Jeff 17 January 2017 / 2:04 am

    after reading peoples story I think I may have had a good result…i am adopted ..I have never searched for my parents before..through out 99 bucks at ancestry ..and bam yesterday my results….1 first cousin 6 second cousins and 19 3rd cousins..figuring out what this might mean but sent out emails to 2nd and 1st cousin matches..nothing yet..but it’s been 1 day..

  15. Elijah Jacob Shalis 11 February 2017 / 5:11 pm

    I can tell you that it has happened to me twice with 2nd half cousins on both my moms side and dads side. Keep in mind they are half 2nd cousins. I know the parentage is right because I have 39 DNA hints and matches on both sides confirming it. On my dad’s side my great grandma had two wives and I descend from the first and they all descend from the 2nd, he had divorced. I don’t match my 2nd cousin Lenny once removed. On my mom’s side my great great grandpa had two wives because the first one died and he remarried and they descend from the first and I descend from the 2nd. I don’t match my 2nd half cousin Dennis.

  16. Colin Alsbury 6 July 2017 / 4:16 am

    For reference I have a group of three siblings (solidly DNA matched to each other) who are related to a group of three siblings (solidly matched to each other) as third cousins once removed. At that level just one of the nine pairings shows no significant matching DNA on Ancestry’s algorithms whilst other matches range up to 4 segments and to over 40cM of match.

    No-one may have ever demonstrated this for second cousins, but third cousins once removed is certainly demonstrable… any bids for any closer relationships??

    As regards more distant relationships I suspect that many people have not researched all possible lines for both parties and that they may find a second or even third relationship between the two people – or there may have been significant endogamy within the tree…
    I certainly have folk that I’m twice or thrice related to where I can’t say which part of a DNA match belongs to which relationship yet without corroborating matches with other people.

  17. Marci 16 October 2017 / 6:56 pm

    How about 2nd cousins 2x removed?

  18. Jaime 26 October 2017 / 10:31 am

    Odd ? , Does Any1 know, if FrenchCanadien ( province of Quebec) are known, for imbreeding? I know, its kind of gross, I’m just curious, it could/may explain more of my weird lineage!! Thanks!!

  19. Gerrard W Rudmin 27 October 2017 / 12:55 pm

    How about 3rd cousins…. is it possible that no appreciable amount of shared DNA will be found?

  20. Eric Yendall 5 December 2017 / 10:43 pm

    I would really appreciate your assessment of this. My mother was born illegitimate. I have tracked-down who her father, my grandfather, might have been. I have been in contact with with the grandson of that person. We have both taken the Ancestry dna test but do not match. Is that conclusive evidence that we are not related?

Leave a Reply

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