I have a DNA match on Ancestry.com that shows a possible 1st or 2nd cousin match at the level of 684 cM's (just barely inside the 680 - 1150 range) and across 29 strands of DNA for 1st cousin. There has been a suggestion that my father is actually the grandfather of the match person.

At that level is that relationship even possible?

2 Answers 2


Please refer to the article The Limits of Predicting Relationships Using DNA. The 684cM of shared DNA precisely complies with the Group C or with the Group D with a little skew. So it is very probable that your guess is correct. But to prove it, of course, you need more tests, because single test can't distinguish who is uncle and niece/nephew, for example. It doesn't give clues about age or generation of person. Just only the distance between two persons to most common ancestor.

If you are interested please write more information: what is the sex of tested persons (I believe that you are man) and from whom you could get additional tests. Also it is very interesting if there are Y-DNA haplogroup and mtDNA haplogroup data in your and this person's test. It may give some additional ideas about relationship.

  • Thanks for the article. So I am a man and actually I am working with 2 people who are 1/2 sisters sharing a common mother. The 684 cM was for the older of the two, the younger is at 461, They are matching with recent tests while my results are from a test done a number of years back, before I ever heard about haplogroups and mtDNA. I am in the process of getting retested. Would you recommend they do so also, and if so using the same provider (Ancestry) or a different one? What we are trying to do obviously is get to the common ancestor and define our actual relationship. Thanks.
    – TommyJoeK
    Commented Oct 3, 2017 at 1:40
  • There are two or three major test companies. They are: 23andMe, FamilyTreeDNA, AncestryDNA. They all provide different type of tests: autosomal, mtDNA and Y-DNA. In case of FTDNA you can choose what particular test you want to buy. 23andMe does not allow to do it: you get all three or none. So you may choose that company what you like more. But the good thing is to download your autosomal data to GEDMATCH service. It has more features for comparing different test sets of data then internal comparing tools of competitors. Also GEDMATCH is the most comprehensive database of autosomal tests.
    Commented Oct 3, 2017 at 8:19
  • Thanks for the link to the DNA Geek article, I've been looking for something exactly like this. I'm puzzled, though, about your comment that "GEDMATCH is the most comprehensive database of autosomal tests". Where might one find figures regarding the number of GEDmatch samples vs. AncestryDNA, 23AndMe and FTDNA?
    – cleaverkin
    Commented Oct 3, 2017 at 19:43
  • @cleaverkin sorry for confusing. I meant that it is "most comprehensive free and independent" database, because it accepts all results from FTDNA, 23andMe etc. Of course, you could estimate the size of database with the open sources like familytreedna.com/why-ftdna.aspx or use the approach as described in the article: thednageek.com/…
    Commented Oct 4, 2017 at 6:40
  • @GEORG GAAL thank you for another great link. Yes, expressed that way, GEDmatch is indeed the most comprehensive database. I would argue that, for better or worse, AncestryDNA is the largest by (to quote Michael Lewis) "a huge, huge, laughably huge margin". Still, you'd need to upload to GEDmatch to get the most useful results.
    – cleaverkin
    Commented Oct 7, 2017 at 6:22

To follow up on the earlier answer -- one of the problems with using AncestryDNA is that there is no chromosome browser. It's difficult to determine what the relationship is from the raw numbers alone -- you have to use that information along with other pieces of data in order to solve the puzzle.

Compare your situation with the scenario recently published in Judy G. Russell's post, "DNA doesn't lie!" on her blog The Legal Genealogist (posted Oct 1, 2017).

Referring to Blaine T. Bettinger's keynote address at the Professional Management Conference of the Association of Professional Genealogists, she says:

Blaine ... showed an example of a case where each and every one of us, looking at the result, would have concluded that the two people tested were parent and child. The one-to-one comparison of DNA at GedMatch showed 3586.7 cM of DNA in common and a 1.0 generation difference between the two.

The two people were aunt and a niece. The missing information: the aunt was the twin of the niece's mom -- making it appear that the two tests were from parent and child.

She also says:

[Blaine] reminded us that DNA alone can never be enough to prove a genealogical relationship. There’s got to be at least one more piece of information to be able to properly interpret the DNA evidence you get.

In a comment to the earlier answer, you said:

I am in the process of getting retested. Would you recommend they do so also, and if so using the same provider (Ancestry) or a different one?

To make an informed decision, you could try learning more about the individual tests given by the different companies. Some resources that might help are:

Think of what tests exist and how they might solve your problem. Y-DNA and mtDNA tests allow males to trace the top and bottom lines of the standard sideways pedigree chart -- see the illustrations accompanying the article on Mitochondrial DNA tests at the ISOGG Wiki. Both women and men can take mtDNA tests, but what you'll learn is only the line from mother to child (men receive mtDNA from their mothers but don't pass it on). Autosomal tests like AncestryDNA give you information that could come from anywhere in the pedigree.

mtDNA tests might add more information, but before you spend the money, make sure you understand what kind of information the test will show you. The article MtDNA testing comparison chart on the ISOGG wiki also includes links to their articles Before You Buy and Choosing a DNA testing company.

Another way to get an idea of whether a particular relationship is reasonable or not is to examine the case studies of other genealogists using DNA as part of their analysis, such as Dr. Thomas W. Jones's presentation "Systematically Using Autosomal DNA Test Results to Help Break Through Genealogical Brick Walls" which was presented as part of the BCG's Quality Day of Education on 6 October 2017. The webinar was hosted by Legacy Family Tree Webinars and the case study will be published soon in an upcoming issue of the National Genealogical Society Quarterly (NGSQ).

In his presentation, Dr. Jones talks about how the amount of DNA people share can vary from the expected percentages and how the apparent distance from the Most Recent Common Ancestor (MRCA) can vary, because of the way DNA is passed down.

  • 1
    All of these comments have been very helpful, and the article references excellent. Thanks so much.
    – TommyJoeK
    Commented Oct 6, 2017 at 21:36
  • I've added a link to Dr. Tom Jones' presentation from earlier today. The recording and syllabus are free through October 13, 2017; after that, they go behind the paywall and will be available for purchase or to subscribers to the webinar series. You'll see an example of using DNA in conjunction with paper-trail research, and what you can learn if you have access to tools like the ones used on GEDMatch.
    – Jan Murphy
    Commented Oct 6, 2017 at 23:25
  • 1
    Thanks Jan! This was really interesting and helpful. And not just in the context of the specific issue I have been pursing.
    – TommyJoeK
    Commented Oct 9, 2017 at 16:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.