DNA: was he or wasn’t he?

Was he or wasn’t he?

It started out as a typical American story. Mom, Dad, three sons and finally a daughter. But when Mom and Dad threw in the towel on their marriage two years after the daughter’s birth, Dad announced that that was one thing he wasn’t — the dad, that is, of the daughter.

Daddy or not?

Now, a reader reports, the family would like to know the truth: was he or wasn’t he the father of the youngest child? Both Mom and Dad are gone now. The daughter is grown and married with a son of her own; her three older brothers are all alive and there’s no reason to suspect any of them might not be the child of both Mom and Dad. So, the reader asks, is there any way a DNA test can give this family the answers they want?

The answer is yes. And it’s actually a pretty easy proposition these days with autosomal DNA testing. That’s the kind of DNA testing that works across gender lines so you don’t have to find a direct male line from father to son to son (YDNA or Y-DNA1) or a direct female line from mother to daughter to daughter (mitochondrial DNA or mtDNA2) — neither of which is an option in this case.3 It tests DNA from all of the chromosomes except the gender-linked X and Y chromosomes,4 and produces information that — used in this situation — can distinguish full siblings from half-siblings.

That’s because of the way autosomal DNA is handed down. Let’s take the case of two people who are believed to both be the children of the same parents. Both of these people will have gotten 50% of their DNA from their father and 50% from their mother. Only in the case of identical twins — who develop from a single fertilized egg — will they both have gotten exactly the same 50% from their father and the same 50% from their mother. Their DNA, then, is completely identical.

With any other sibling pair — even fraternal twins5 — both will still get 50% of their DNA from their father and 50% from their mother, but it won’t be the same 50% from either parent. That’s because of what’s called recombination, the random shuffling of DNA segments that goes on in each generation.6 On average, then, any non-identical sibling pair will share about 50% of their total DNA. That’s roughly the same amount of DNA each person has in common with a grandparent or an aunt or uncle.7

Also by random chance, each ordinary sibling pair will have some areas in their DNA where each of them did get the exact same segment of DNA from Mom and Dad that the other got — say, segment 1 from Mom and segment 1 from Dad. Those segments are called “completely identical by descent.” They’ll have other areas where they both got the same segment from one parent but not from the other — say, both got segment 1 from Mom but one got segment 1 from Dad and the other got segment 2 from Dad. Those are called “half-identical by descent.”8 And, of course, they’ll have other areas where they just don’t match each other at all because one got segment 1 from Mom and segment 1 from Dad in that area and the other one got segment 2 from Mom and segment 2 from Dad.

Now if the siblings are half-siblings — if they share only one parent — these results will change in two ways. First, on average, half-siblings share less overall DNA than full siblings, 25% as opposed to 50%.9 Second, while half-siblings will have some segments where they don’t match at all, and some segments where they’re “half-identical by descent,” they won’t have any segments that are “completely identical by descent” since that’s only possible where those tested share two parents.

So if any one of the brothers in this family’s situation will agree to be tested along with the sister, they will be able to determine if they are full siblings… or not. There’s no benefit to also testing the sister’s son — his DNA results won’t add to this picture at all.10 Testing more than one brother can’t hurt, except in the pocketbook. And I’d definitely go with autosomal testing rather than what’s called sibling or kinship testing because it tests more of the overall DNA.11

Generally, for DNA testing for genealogical purposes, I tend to recommend the Family Finder autosomal DNA test from Family Tree DNA because it’s designed by genealogists for genealogists and you tend to get more information from cousins you discover through the testing process. But this is one situation where I’d recommend the Relative Finder autosomal DNA test from 23andMe. And I do so because 23andMe has one option, called Family Inheritance : Genome View, that displays the results just the way I’m describing them here: you can actually see whether two people have any segments that are completely identical (they share both parents) or not (they share only one parent).

And, of course, if this family decides to go on and use the testing for their family history research, they can transfer their 23andMe results to Family Tree DNA and use both systems — the best bang for the DNA buck.12


  1. ISOGG Wiki (http://www.isogg.org/wiki), “Y chromosome DNA test,” rev. 23 Jul 2011
  2. ISOGG Wiki (http://www.isogg.org/wiki), “Mitochondrial DNA,” rev. 30 Jul 2010.
  3. A daughter never inherits a Y chromosome so she can’t be tested against her brothers’ Y-DNA, and the family already knows that they all have the same mother, so there’s no reason to test the mtDNA.
  4. ISOGG Wiki (http://www.isogg.org/wiki), “Autosomal DNA,” rev. 8 Feb 2012.
  5. Fraternal twins are “twins, not necessarily resembling each other, or of the same sex, that develop from two separately fertilized ova.” “Fraternal twins,” Dictionary.com (http://dictionary.reference.com : accessed 23 Jun 2012).
  6. ISOGG Wiki (http://www.isogg.org/wiki), “Recombination,” rev. 30 Jul 2010.
  7. See generally Judy G. Russell, “Autosomal DNA testing,” National Genealogical Society Magazine, October-December 2011, 38-43. See also See Ask a Geneticist, posted 10 Oct 2008, The Tech Museum, Stanford University (http://genetics.thetech.org/ask-a-geneticist : accessed 23 Jun 2012).
  8. See generally ISOGG Wiki (http://www.isogg.org/wiki), “Identical By Descent segment,” rev. 20 Aug 2010. Note that ISOGG uses the terms “Half-Identical Region” and “Full-Identical Region.”
  9. See Ask a Geneticist, posted 25 Jan 2012, The Tech Museum, Stanford University (http://genetics.thetech.org/ask-a-geneticist : accessed 23 Jun 2012).
  10. He won’t share any Y-DNA with the brothers because he’ll have gotten his Y-DNA from his own father, the sister’s husband. And the mtDNA is already known.
  11. See Richard Hill, DNA Relationship Testing Checks Kinship, DNA Testing Adviser (http://www.dna-testing-adviser.com/index.html : accessed 23 Jun 2012).
  12. Judy G. Russell, 23andMe to FTDNA? YES!, posted 5 Feb 2012, The Legal Genealogist (http://legalgenealogist.com : accessed 23 Jun 2012).
Print Friendly
This entry was posted in DNA. Bookmark the permalink.

12 Responses to DNA: was he or wasn’t he?

  1. Jenna says:

    Thank you Judy, you’ve explained that so this DNA testing wanna be can understand the differences. I also appreciate that you explained why you prefer Family Finder.

    • Judy G. Russell says:

      So glad I could help, Jenna! Let me just say for the record: I’ve yet to meet a DNA test I wouldn’t take and I’ve tested with 23andMe as well as Family Tree DNA (and, finally, now, I’ve gotten off my duff and sent my sample off to AncestryDNA as well).

  2. Anne Willson says:

    hmm….I’ve got one of those exact problems. Grandma shows up in 1920 Census saying she’s still married to Grandpa. But her brother shows her married to the second husband on his 1917 WWI Draft Registration (same address in both cases). My aunt was born 1914. Sounds like a great plan for checking…quick before anyone else dies!

  3. This is really thorough! A clear explanation. You must have a legal mind. ; – }} I’m retweeting this for reference in case my cousins are willing to take a DNA test after we have our family reunion discussion. Thank you, Judy!

  4. Carolyn Lea says:

    Thank you – I am learning a lot from your posts on DNA as well as reading Debbie Kennett that was recommended, Blaine Bettinger, Debbie Wayne Parker, and information from ISOGG.

    Your posts are very clear – the above helped give me a better understanding of Autosomal DNA and why I should go ahead and have my brother do FF and YDNA for xmas. Even with the sale at FTDNA I can only afford 37 marker if I do FF. Being Jewish on my father’s side I was hoping that YDNA might lead to matches prior to surnames at some point and give more information on which of the my FF matches actually may mean something as I know being Ashkenazi is an issue.

  5. Lucinda says:

    Nice article. I wonder, though, about the last sentence in paragraph five. Grandparents/grandchildren, uncles or aunts/nieces or nephews, and half siblings all share about 25% of their total DNA, statistically. If you refer to sharing identical + “half identical” regions after DNA testing, the sentence would be correct. Parents/children and full siblings share about 50% of their total DNA, statistically, but 100% of their identical + half-identical regions after testing (but not necessarily the exact same regions). Just to clarify.

    • Judy G. Russell says:

      You’re right and the sentence was pasted into the wrong paragraph. It should be in the seventh paragraph referencing half siblings, not full siblings.

  6. Lucinda says:

    Actually, only full sibs will likely share identical plus half-identical regions across their genome, but still only share at 50% of their DNA. Half-sibs will not have identical regions.

    • Judy G. Russell says:

      That point is made and clearly so: half-siblings “won’t have any segments that are “completely identical by descent” since that’s only possible where those tested share two parents.”

  7. Lucinda says:

    That makes sense! Sounds like something I’d do. Feel free to fix and delete my comments.

Leave a Reply

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

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>