Caution: Invisible Fathers and Autosomal Matching – Who’s Hiding in Your DNA?

caution Y

Using autosomal DNA matching alone, NPEs (nonparental events,) also known as misattributed paternity or parentage (MPEs,) undocumented adoptions, and sometimes other terms, often goes completely undetected, especially a few generations back in time.

Generally, this phenomenon occurs when the believed father is not the biological father – for a variety of reasons. Using autosomal DNA alone, especially more than a couple of generations back in time, these situations often go undetected because we are lulled into complacency thinking we have proof via DNA matching, but we don’t.

Let’s look at a real life example of how I discovered and unraveled one of these mysteries.

What We Thought Happened

In the Doe family, two male Doe men were found together in King and Queen County Virginia during the Revolutionary War era. They moved together to Halifax County after the war. They lived very near one another and witnessed documents for each other, lifelong. They were known within the family to be related.

Thomas was older than Marcus, but not enough older for Marcus to have been his son.

Eventually Thomas wrote a will that Marcus witnessed, but Marcus was not mentioned in the will. Typically, unless facing imminent death and the dying person created a spur of the moment verbal will in the presence of whoever happened to be in the room, called a noncupative will, family members who benefited from the will did not witness the document. In probate records, noncupative wills were stated as being such – and Thomas Doe’s wasn’t.

Autosomal NPE1

Enter the age of DNA testing. Many of Marcus’s and Thomas’s descendants match each other autosomally, distantly of course, as would be expected.

Everything was just hunky-dory until a Y DNA test of Marcus’s patrilineal descendant (male tester Doe) and Thomas’s patrilineal descendant (son Doe tester 2) proved conclusively that the Y line DNA was NOT the same in those two branches of the family.

Furthermore, Alexander Doe actually had another son, James, whose sons’ descendants Y DNA match those of Thomas, so the Thomas/James Y DNA proved the Y DNA of Alexander, leaving Marcus out in the cold. Furthermore, James was near Marcus’s age, so he could not have been the father of Marcus.

Clearly, Marcus, Thomas and James were related, because the autosomal DNA of their descendants matches and triangulates.

What gives?

What Really Happened

When Alexander Doe died, with a will, his eldest son, Thomas was made the guardian of his three underage daughters. One married a few years later, but the rest of the daughters disappear into the mists of time.

Adding to the confusion, the King and Queen County, Virginia, records burned, destroying what evidence might have existed. Fortunately, the family lived on the border with Essex County and a few records were filed there.

autosomal NPE2

I’m giving you the answer to the question of what happened now, because I want to discuss how this misattributed parentage went undetected, and why you CAN’T detect it without Y DNA testing.

Marcus wasn’t the son of either Alexander or Thomas, but was the illegitimate son of Susan, daughter of Alexander, when Thomas was her guardian. Susan probably died, because given the nature of how close Thomas and Marcus were, Thomas probably raised Marcus and Susan is never mentioned again.

Of course, the autosomal DNA can match because autosomal DNA does not distinguish between DNA descending from maternal or paternal lines, like Y or mitochondrial DNA, it only looks at matching.

The only way to ever “notice” that Marcus was NOT patrilineally a Doe was to Y DNA test Marcus’s male descendant and a Doe male descending from Thomas Doe or brother James.

Due Diligence

If you’re asking how we know that the break in the Y DNA wasn’t between Marcus and the current male tester in his line, that’s a great question and shows that you’re thinking.

You’d need to Y DNA test a male descended from Marcus through a different son. Of course, in this case, we already have two sons of Alexander whose Y DNA matches so that line is proven.

Why We Can’t Tell Using Autosomal

Let’s look at the original presumed relationship again – the relationship reported in all of the family history documents written in the early 1900s. This is what we believed, of course.

autosomal NPE3

Using the chart I compiled from Blaine Bettinger’s Shared cM project, here, I added the columns above reflecting the relationship between Marcus, Susan and Thomas if they were all children of Alexander. You can see that the various testers today were 6th cousins, or Degree 13 relatives who would be expected to share 21cM of DNA on average. The range of shared DNA for 6th cousins is from 0-86 cM. These cousins all shared in the expected range for this relationship, so no eyebrows were raised.

autosomal NPE4

Looking at the real relationship in the chart above, you can see that the testers are actually “once removed” meaning that one is a generation further down the tree than the other one. This makes them 6th cousins once removed, or Degree 14 relatives who share 16 cM of DNA on average. However, the range of DNA seen at 6C1R is 0-72 cM, still very close to the Degree 13 relationships. The two men in question, as well as the rest of the descendants of Marcus, Thomas and James all match within the expected range for either 6C or 6C1R.

The bottom line is that without Y DNA testing, one can NEVER assume (you know about that word) the line of descent is as presumed, or even as documented in records, even with an identical-by-descent autosomal DNA match. (To read about identical-by-descent versus identical-by-chance, or false matches, click here.)

An autosomal DNA match (at this distance) only shows THAT the two people are related, not HOW they are related.

Records have been known to be incomplete or wrong, intentionally or otherwise. We need combinations of records and multiple kinds of DNA tests.

Y DNA to the Rescue

Without Y DNA testing, Marcus would forever have been believed to be the son of Alexander. Some researchers attributed Marcus as the son of Thomas, because of their close relationship in Halifax County. Now we know that he was unquestionably not the son of either man, and thanks to Y DNA we know the likely candidate surname of his father thanks to multiple high-quality Y DNA matches.

In time, utilizing autosomal+Y DNA combined with genealogical records, we may actually be able to narrow down the candidates for Marcus’s father or at least candidate families. Susan had to be geographically close to Marcus’s father to conceive Marcus, and neighbors are always the best candidates.

Advanced Searching for Combined Matches

Another benefit of Y DNA testing is that Family Tree DNA offers a combined matching tool. From the account of Marcus’s descendant who carries his Y DNA, we can query the system for anyone who matches both the Y DNA and autosomal DNA through the Family Finder test.

autosomal NPE advanced

Click to see a larger image. Note the selected options.

This technique provides a way to determine who matches on both the Y and autosomal tests. Of course, there’s no guarantee that any two people will match autosomally beyond 2nd cousins. After that, it’s a bit of a crap shoot, and the further back in time, the less likely people are to match. However, you won’t know unless you try. If you do match on both tests, keep in mind that the dual match can also be from two different ancestors.

However, if the Y DNA matches the known patrilineal line, you won’t have to worry about misattributed paternity because you know the patrilineal link is solid (unless of course another male from the same line was involved.) Y DNA can’t tell you who, exactly, just that you match the patrilineal line.

You need the combination of both Y and autosomal DNA.

Which of Your Lines Need to Y DNA Test?

How many of your surname lines are you assuming descend through the direct paternal line when in fact the match could be a result of a sister or female relative giving birth and giving the child her surname?

Autosomal NPE sister

When illegitimacy is involved, the child takes the mother’s surname but the father’s Y DNA is passed to the male child. The child carries the same approximate amount of autosomal DNA of the grandparents as if the child had been born to the brother in our example, and the family often went to great pains to hide the fact that the birth was illegitimate in the first place. After a generation or two, the ONLY way to know is to Y DNA test.

Who might be hiding in your tree? Which lines do you need to test? Who is available from those lines to test?

You can purchase both the Y DNA and autosomal Family Finder tests at Family Tree DNA. You can order a 37 marker test here which will be enough to see if you match the surname you expect. Be sure to join relevant projects too, especially the surname project. I wrote about how to join a DNA project here and how to determine which projects to join here.


38 thoughts on “Caution: Invisible Fathers and Autosomal Matching – Who’s Hiding in Your DNA?

  1. Thank you! Your chart explains cousins much better than anything else I have seen. Since I have several very distant cousins, this helps a lot!

  2. Thank you for this article!

    Our family began with Y-DNA testing, because we weren’t sure we were really Browns. My gr-grandfather was taken in by a couple who were told that his name was George Brown, but we weren’t convinced.

    When my brother did Y-DNA testing we were able to verify that we were Browns! We joined the Brown Project, and found that we were Brown Group 10, but that alone wasn’t enough to help us find George’s parents.

    When my 90-year-old father did autosomal DNA testing, we found answers! Multiple matches pointed to a specific family, with the closest matches pointing to a specific person – and that family had already been identified as being a part of Brown Group 10!!

    …now we’re trying to discover the identity of George’s mother. We also have multiple autosomal matches going back to a specific couple who were located in the same area as the father; and that couple had a number of daughters…
    : )

  3. What a great article!! I ‘get it’ now! Most importantly, details like this really matter! Thx for this one especially!

      • Great idea. I put my son in this group. I would be really interested in digging further as not much luck with close matches so far. Makes me wonder if there was a little issue somewhere along the line!

  4. While you start your discussion with autosomal dna matchies, it was a surprise result in my Y-dna test that motivated me to be further tested and to have multiple relatives tested for Y, mitochondrial, and autosomal. So far I’ve only been able to confirm the most recent three generations. But that said, there is one point in your discussion I wonder about. Specifically, you say (if I read correctly) that Thomas’ line is confirmed because the Y-dna matches along both Thomas’ and James’ lines. Hypothetically, and in the absence of additional information, is it not possible that both Thomas and James were sired by someone other than Alexander and that Marcus was sired by Alexander? Surely more than one woman has had multiple children by the same man not her husband . . .

    • It’s possible, but less likely that Marcus being the odd one out, especially with the combination of the fact that Marcus is not in Alexander’s will, but both Thomas and James are. Marcus’s age is an issue too for the mother.

  5. Hi Roberta, I love your articles. Can you explain how you might match one surname at 37 markers but not any of that same surname at 67? This is for my male cousin who has done a 67 marker test but doesn’t have any matches to our surname Roberts at that level. Maybe the next article☺

    • In some cases, people don’t test any higher, so watch for that. In other cases, there is just one too many mutations that cause the match to not be considered at the next level, especially if the DNA was very common at the lower levels.

  6. Thanks again for illuminating post. Following your Y-DNA analysis, two Y-DNAs, lets say FTDNA Y111, how similar is the Y111 between two siblings, between a man and his grandson etc. Have you made a table, similar the of of similarity of autosomal along the family tree?.


    • It can vary widely. My Estes ancestor immigrated in 1647 and had many sons. In one of those lines, there were no mutations based on reconstructing the DNA haplotype of the ancestor. In another line, many. It depends on the roll of the dice. I have seen fathers and sons vary by 3 mutations, although that’s rare. The TIP calculator provides statistical estimates between two matching men.

  7. I have had Ydna tested for my father, maternal uncle and husband. None of them match anyone with their surname, I have determined through autosomal matching that my father’s father was from a family with several sons. My uncle and husband I believe are probably much further back and have yet to find someone to ydna test. I find it amazing that none of them have ydna matching their surname. Aren’t I lucky to have so much work to do?

  8. My question was: “absent other information . . . ” was my hypothetical possible and I read your answer as “yes.” The point, then, is that while you have satisfied your intuitive criteria to have “proven” the James/Thomas lines as being from Alexander but without further evidence there is still room for doubt. Frankly, I would take the same position under those circumstances, but I’d want to be very careful that I wasn’t inadvertently creating an implicit bias in my interpretation of the other evidence as it appeared.

    But let me echo the others in thanking you for the article – definitely helpful in framing my thoughts and methodology for my own work.

    • Yes, it would be very difficult to reach 100% certainty, even with perfect records. Even if a woman had 10 male children, could all of those sons been fathered by the same “other man?” Of course, is is likely. No – but it is possible. Without other generations in the same male line, we can never know positively without exhumation. And exhumation, unless the body is in a crypt with a chiseled name, is fraught with the potential of mis-identification as well. For example, I thought my own father was buried on the “other side” of his tombstone. If I had excavated him, I would have been in for a rude awakening when I got DNA results back, not to mention a huge bill to dig up the wrong guy.

  9. It would be nice if more lines y-DNA tested. I administrate the Short y-DNA Project. Before testing, it was thought that Aaron Short b. ca. 1774 may have been a son of Thomas Short (b. ca. 1751) and a brother to Thomas Short, Jr., born ca. 1775. After testing, it was found that descendants of Aaron are an exact 37 marker match for a Grigsby line.

    It is thought that Aaron’s mother was Permelia “Milly” Short b. ca. 1758, a sister to Thomas above born ca. 1751. Permelia and Thomas’ mother Ann (the widow of Samuel Short) secondly married to John Breeding, who had a daughter Winifred Breeding by his first wife. Winifred married John Grigsby in 1778. Descendants of that John have a distinctive marker that is also shared by descendants of Aaron Short. While it doesn’t prove that John Grigsby was the father of Aaron Short, with the distinctive marker it had to be either him, or someone close to him. (Last time I checked there weren’t tests from any of his brothers).

    Court records show that Permelia Short had several children outside of marriage with Richard Thompson in the 1780s and 1790s. It is thought she is also the mother of Aaron Short. If John Grigsby is the father, then he had an affair with Permelia Short about 1773, before he married her step-sister Winifred Breeding in 1778.

  10. Roberta,
    Wow, awesome blog a definitely gives me food for thought especially when it comes to tracking down my Father’s Biological Father and his family. My Grandmother had my Father out of wedlock but they did such a job of covering it up that my Grandmother didn’t tell my Father until he was an adult and had a family of his own. My Father did try to contact his Biological Father but he refused to acknowledge my Father. My Father has passed several years ago and when I did my Brother’s DNA the only surname that popped several times was McCall but on the back of several pictures of my Father that I have the last name was of McClabber…so I know my Grandmother possibly was not honest about that either.
    My Grandfather adopted my Father when he was 3 yrs old but I have been unable to find any records regarding this, but still working that angle…lol. I do know that in the 1940 census I found my Dad at the age of 3 yrs old and the surname of McClabber was not spelled like this but like this M (3 little dots) Calabber, not really sure what the 3 little dots mean but that is how my Grandmother put it on the census .
    Definitely this blog is giving me hope and one day I will find my Father’s biological Father and family.
    If you could answer what the 3 little dots mean that surrounds the Letter M that would be awesome…thanks for this article very interesting and insightful

    Cindy Carrasco

  11. Great article Roberta! And certainly proves the value of Y-DNA testing. My question is how do you know Susan is the mother? Did you do further triangulation of the autosomal DNA results for other descendants of Susan, Thomas and others to determine this?

    • In reality, Susan is the oldest daughter and entirely disappears from the record. Normally, a guardian would disperse funds when the child came of age or died, but not in this case. She disappears. One sister marries. So I’m not actually positive it is Susan, but the only candidates have to be the three sisters. Alexander was an immigrant from Scotland and there were no other men of that surname. The DNA does triangulate, so we know positively it’s the same family line. And Thomas has a very close relationship with Marcus. So, no there is no proof positive that it was Susan and not one of the other daughters.

  12. After a couple of things that showed up in DNA in our extended family, I would say don’t have it done unless you are prepared for some surprises.

  13. Thanks, Roberta, for another thought provoking blog.

    We have a similar situation in one of my Davis lines where our brick wall is Solomon Davis b 1798 in NC. I noticed that he and his family traveled for decades with an older Richard Davis and his daughter Sarah who had a son out of wedlock “C” and he kept the Davis name, although his mother married Austin Stockstill and had more children. Unlike your example, “C” knew he wasn’t a Stockstill.

    When I contacted family members wanting a YDNA test, we realized “C” had his father’s Y-DNA who wasn’t a Davis, so none of his descendants could be our testee. Autosomal testing has tied us to this family, but I surely would like to find a true male descendant of Richard. YDNA 111 tests match descendants of a Benjamin Davis but he’d also have to be a cousin on this side of the proverbial brick wall. The search continues!

  14. We have a situation in our Powell Y-DNA Haplogroup T in which one of my brother’s Powell Y-DNA matches in Haplogroup T descends from a female with maiden name Powell who is the daughter of a known Haplogroup T Powell male ancestor. She had a son with an unknown male who was also a Powell. We know this, because the Y-DNA of his male Powell descendants is also in Y-DNA Haplogroup T and matches other Powells in Haplogroup T. Not sure how to even begin figuring out who this unknown Powell father was. Any ideas?

    There is also another situation in our Powell Y-DNA Haplogroup T in which another one of my brother’s Powell Y-DNA matches has the surname Sizemore, but is a close Y-DNA match of my brothers’ — genetic difference of 2 out of 111 markers. Our line of Powells has been proven through Y-DNA and supporting Autosomal DNA to descend from Thomas Powell (ca.1640-1701) and Mary Place (ca.1649-1710) via one of their sons — possibly Honorias Powell (as my 6th g-gf named one of his sons Honorias “Honor” Powell. Note: the name Honorias is unique to our Haplogroup T Powell line, as new Will evidence has come to light that indicates the first Honorias was named after his mother Mary (Place) Powell’s mother (or possibly stepmother) Honoria “Honor” (MNU) Place Weir Jones.

    Our branch of this Powell line descends from a William Powell (ca.1725-1791) from Virginia who settled in Bute Co., NC — which became Warren Co., NC. I found in land records a William Powell and a William Sizemore owning land beside one another in Amelia County, VA in the 1750s. I also found in Amelia County, known descendants of our Powell line — Honorius and Simon — referred to in estate records. The William who lived beside William Sizemore in 1750s Amelia County, VA was referred to in land records as William Powell of Caroline County. I read some abstracts of court documents for Caroline County dated before that time which stated that a William Powell, Henry Powell and others had their land in Caroline County, VA taken away from them because they voted for a Mr. Gibson instead of Mr. Martin in a VERY contentious election. When Martin won, he got revenge by taking away their land. I suspect that may be how they ended up in Amelia County, VA. Perhaps they left to start over somewhere else.

    Our Powell line has been proven through land and estate records to have lived in Caroline County, VA at one time. I’m not sure whether the William Powell of Caroline Co. who also lived beside William Sizemore in Amelia County is our ancestor and if the William Sizemore who lived beside William Powell is the ancestor of my brother’s Y-DNA match Mr. Sizemore. However, it seems reasonable that he may be. I’m just not sure where to go from here. Any ideas on this one would be appreciated as well.

    — Deborah

    • On the second situation, have you looked for triangulatiing autosomal segments? The first situation is a real humdinger. Makes you wonder what happened.

      • I see a couple possibilities for that first scenario – a 1st cousin marriage is the simplest and most palatable. Another, while it isn’t the most comfortable hypothesis, could be through incest – father/daughter or brother/sister (more likely the former). I ran into that situation last year when trying to explain why one of my distant matches, whose father had been adopted at birth and trying to find his bio family, matched others at a number that significantly exceeded Blaine’s range for his expected relationship. It was eventually confirmed with knowledgable family members. Proving either possibility five or six generations ago would likely be near impossible without additional evidence.

    • I match “your” Henry William Sizemore kit ftdna 3213, he did a BigY and placed it in Yfull …YF06979 Henry Sizemore b. 1810 and d.1977…….He is Ydna T1a2 branch the same as your Powell who is T-L131 , again T1a2 branch.
      You need to triangulate with the Hill families who are also T1a2 branch
      David A Hill 1850, ……..Georgia
      Nathaniel Hill 1825, …….Virginia
      William A hill 1782 …North Carolina
      and John M Hill 1819…South Carolina

      I am also T1a2 , but from the Italian Alps, Veneto and South Tyrol regions…….my match with this Sizemore must be at least over 1000 years ago.
      Sizemore was a saxon family , known as Sigmar

      contact me if you like to discuss

  15. “…Furthermore, Alexander Doe actually had another son…”
    1: Did I miss Alexander being introduced to the scenario.
    2: ‘Another’ son? Where is mention of his first?

  16. Thank you. Just like with regular research, the more ways you make a connection, the more accurate the case is. And in this case, it helps point you in the proper direction for figuring out where to look for records.

  17. One story and one question

    I have y-dna conundrum I still haven’t solved (i’m putting off a research trip to the Virginia archives in Richmond and Chesterfield because it will be a full day event and I don’t know when I will have the time for that). The hardest branch to research has been my father’s direct patrilineal line.

    Doing a y-dna (y67) test has helped me because 1) I discovered there are 4 or 5 different ydna lines from Virginia with the same surname and same alleged family group (1 native american ydna line, 2 different r1b1 lines, a possible 3rd different r1b1 line, and another family that hasn’t ydna tested who may or may not match one of the r1b1 lines) 2) i discovered my ynda traces to northern England and Southern Scotland 3) I match an unexpected ydna subgroup. None of those men in my ydna subgroup match me as a cousin in an atdna test.

    My farthest known direct male ancestor had two sons, John and Jeremiah. Most of my male cousins can trace back to John. I trace back to Jeremiah. We believe them to be full brothers. I really need one of my male cousins descended from John (Jeremiah’s brother) to do a ydna test. Hopefully, they match me. If not, well, that would just make things more complicated, wouldn’t it?

    And now my fun question, and it relates to atdna testing. My brothers tested at ancestry. He has a match, that we call Person A. Person A matches at about 12cM and shares no other matches with my brother at ancestry. Sounds like a false positive or a one off, right? Person A transferred their results to gedmatch. Person A now matches my brother and I at approximately 45cM, and now they belong to a very specific family cluster.

    Is this one of the strange results of imputation? I’ve asked the dna match to consider transferring their dna results from Ancestry to FTDNA. Better resolution might be achieved if they tested separately at FTDNA.

    Any thoughts? I haven’t asked Gedmatch (yet).

    • GedMatch is struggling right now. Yes, have them retest at FTDNA. Also, at a certain level, Ancestry simply doesn’t report shared matches – it’s not that you don’t have them. You may have hit that threshold.

  18. Pingback: Y DNA Resources and Repository | DNAeXplained – Genetic Genealogy

Leave a Reply