Genetic DNA test turns up a surprise... what should I do?

November 12, 2013 8:46 AM

I am a male. I am the family historian. In an effort to fine-tune my own research into my maternal grandfather's family (the "Jones"), I have my male first cousin take a Y-DNA test, since he has the "Jones" Y-Chromosome (I have my Dad's Y chromosome). Lo and behold, the test comes back with a perfect match, and a predicted Most Recent Common Ancestor within 3 generations. I have researched back farther than that and I know all of their names. None of the last names match. My occupation doesn't take me anywhere near genetics... but I am assuming that a Y DNA match with a MRCA that close should have the same surname for his male ancestor, correct? Except if there's been adoption or infidelity or rape, correct? I sent the results to my cousin, prior to the lightbulb going off over my head. He knows a lot less about this than I do. He hasn't called me yet, but what if he does? Should I just let sleeping dogs lie? Ugh!
posted by anonymous to Science & Nature (23 answers total)

You may seeing something more sinister than the truth. 3 generations is a while and despite what we like to think about the past, people got remarried and or had kids out of wed-lock a lot. Without infidelity, rape or incest being involved. This is especially common in smaller towns.

I would not worry about him having a "bad" reaction to this.
posted by French Fry at 8:57 AM on November 12, 2013 [3 favorites]

people got remarried and or had kids out of wed-lock a lot

My mom did the ancestry.com thing a few years back, and one of the more interesting things we learned is that my great grandparents got married a year later than the family thought they did, meaning my grandma was born something like 2 months after they got married rather than a year and 2 months. This fact had been quietly glossed over my grandma's whole life.

So yeah, I'm on team Probably Nothing Nefarious Happened Here. Born out of wedlock, remarried, bad records keeping, kid lived with a family member other than his parents, immigrant surname whitewashing, etc, etc. Any number of things could have happened to change the last name, not necessarily something traumatic.
posted by phunniemee at 9:07 AM on November 12, 2013 [3 favorites]

Agree with the others that it may not be anything particularly nefarious, and in the unlikely event your cousin does ask you about it, I don't think you owe him anything more than, "I know, that's weird, isn't it! I wonder what that's about."
posted by mskyle at 9:09 AM on November 12, 2013 [3 favorites]

It is difficult to parse your question. Are you saying that your cousin is called "Jones" but that nobody further back than his (and your) grandfather was a Jones? Or are you saying that your cousin has a different name from his grandfather and you don't know where it came from?

Either way, there are loads of ways that names can suddenly appear in the family history. Many people - especially in the period when your maternal grandfather would've been around - change their name for all kind of reasons, often to do with migration etc.

My girlfriend's mum, back in the 1970s, decided that she didn't like her surname (it is very common) and gave herself an entirely new name (a totally made-up word). When she got married, her husband took the name too, and later other family members adopted it as well. It's now my girlfriend's name, and it's likely that it will be at least part of the names of any children I have. In a few generations, anyone looking at the family history is going to be totally stumped about where this word came from. I'd be disappointed to imagine that they will jump to the conclusion that rape or incest was involved.
posted by cincinnatus c at 9:18 AM on November 12, 2013 [5 favorites]

I am with the responses above who say there is not necessarily a scandal at the root of this. I learned some time in my adulthood that my great-grandmother Biscuit -- that is, the woman I understood to be my paternal grandfather's mother and whom I remember faintly from my childhood -- was actually not among my direct ancestors at all. Great-Grandpa Biscuit married a woman in 1918 and they had four kids together. When the kids were all very young, his wife became ill and her sister moved in with the young family to help out. Eventually the ill wife and mother died and after an appropriate period of mourning, Great Grandpa remarried, to his erstwhile sister-in-law.

This is not especially well-known in my family, and I learned it only when talking with a relative about some genealogical trivia. I have a dozen or fifteen siblings and first cousins on that side of the family, and I'd be surprised if more than one or two of them know that our respective parents' Grandma Biscuit had previously been their great aunt.
posted by ricochet biscuit at 9:23 AM on November 12, 2013 [1 favorite]

I think what's going on is that Anon and Jones are related on Anon's mom's side, and as far as he can tell, Anon's dad doesn't share any known relatives with Jones, but according to their Y-DNA test they have a relative in common on their direct paternal line relatively recently. This doesn't seem, to me, to be the case of someone changing their name - it's an issue of Anon and Jones actually sharing a grandfather (or Anon's grandfather being Jones great-grandfather, or some combination like that) without being reflected in actual records.

I think all you can really say is that there's a mystery here. It seems to me like you're assuming that the mystery is on Jones's side, not on your side, but right now that's equally likely. Have you been asking other male relatives to get YDNA tests? That might help to start clearing things up. Either way, discovering mysteries like this is one of the risks of DNA testing.
posted by muddgirl at 9:31 AM on November 12, 2013 [5 favorites]

Chiming in to agree: there is likely no Deep, Dark Secret afoot, just something that might have been regarded as scandalous back in the day and isn't now (like divorce or out-of-wedlock pregnancy) or not made clear in family history (a parent who is really a step-parent, like Ricochet Biscuit's example).

Ancestry.com is helping to uncover this sort of family "secret." I'll add my own example, which is really close to Phunniemee's: My grandma always said that her mother's parents (my great-great-grandparents) had married back in the old country and arrived together in the US along with an older sister of my great-grandmother's, who died as a baby.

When I researched that line of the family on Ancestry, lo and behold: my gg-grandparents had met in the US, married there, and my great-grandma, their firstborn, was born some six months after their marriage. "Married in the old country" and "dead older sister" was something my grandma invented to cover up that shotgun marriage in the family tree. I had a good laugh when I found out what happened. People in those days had no idea that their carefully-buried family secrets would be uncovered by Ancestry.com.
posted by Rosie M. Banks at 9:36 AM on November 12, 2013 [5 favorites]

I've just started getting into using DNA for genealogical research, so I'm not an expert by any means. However, I have heard that you should take the generations that these services mention with a grain of salt. Somewhere I've heard to add 2 -- so if they say 3, it may be 5. I've come up with loads of connections and have only pinpointed the actual connection with one person. I believe GEDMatch said 4 generations, but it was really 6.

I've also been unable to find a match with many connections and I've researched my family pretty far back pre-DNA testing.
posted by jdl at 9:37 AM on November 12, 2013 [3 favorites]

There was also a lot of trading of kids around in the past (just like today, but without the adoption paper trails). My mom was an only child, but she grew up during the Great Depression with one or two other kids that my grandparents took care of. In a more caring family than my mom's, those kids would have been welcomed in with last names changed and all the benefits and responsibilities of being part of the family. You (or your cousin) could easily have a grandparent or great grandparent that isn't a blood relative to the family that they grew up in with nothing more nefarious going on than one of the several economic depressions that happened and tore families apart between 1880 and 1940.

Also, humans like to pretend that they are faithful to one spouse for life, but reality is that people have always slept around, the definition of marriage and fidelity have changed over the years while always being the 'official version' rather than what real people are doing in the real world.
posted by BearClaw6 at 9:41 AM on November 12, 2013 [1 favorite]

I can't add anything regarding the question of whether or not there may have been a scandal in your family's past; this sort of family tree stuff confuses me.

However, assuming the worst, and if there were some sort of anomaly surrounding your cousin's origin's, why is this a problem for the two of you? It's a part of your family's shared history. It's a part of your and your cousin's past.

I think it's okay to talk about and you don't need to be afraid.
posted by univac at 9:43 AM on November 12, 2013 [1 favorite]

I think what's going on is that Anon and Jones are related on Anon's mom's side, and as far as he can tell, Anon's dad doesn't share any known relatives with Jones, but according to their Y-DNA test they have a relative in common on their direct paternal line relatively recently.

That sounds like it too. But this question is still hard to parse. Is the MRCA calculated from the Y-DNA or are you talking about the MRCA in general? The latter, of course, is not surprising since you two are, in fact, cousins.

Also what exactly is a "perfect match"? Do you mean you have the same haplogroup (not very interesting) or that your Y-DNA matches exactly (doubtful)?
posted by vacapinta at 10:15 AM on November 12, 2013 [1 favorite]

Which test (no autosomal?), which company, and how many locations were measured? You'll get different numbers of matches on different sites for different numbers of markers. For example:

FTDNA: 12 markers = 128 exact matches; 25 = 23; 37 = 1; 67 = 0. That match on the 37 marker test is a fifth cousin, FYI.
ySearch: 8 = 22 (one is me); 30 = 3 (one is me).
Ancestry: 24 = 3; 30 = 2.

In the early days I'd get messages from matches on the lower marker tests all the time and never found anything close to a link with them.The 37 marker match at FTDNA is my lone success, though the 2 extra folks at 30 on ySearch are from the same general area as I am from, so we are probably related from gobs of years ago.
posted by jwells at 10:54 AM on November 12, 2013

No, no, no, no, no! This is all wrong. Don't go drawing conclusions about your family from a test like this.

I'm a geneticist and I had some trouble understanding your question. However, one thing that's very important to know before drawing conclusions is how common the marker you looked at is, in your population.

You won't have looked at the whole Y chromosome, just a tiny piece. If it's very common in your population, it wouldn't be a surprise that your cousin shares it ("a perfect match" - though I couldn't work out with what?) with lots of others. I work with MRCA analysis all the time - it applies over much longer time scales (tens of thousands of years at the very minimum) not to pedigree-level analysis within families (surnames and generations), so I also didn't understand that part of your question.

{And that's even before we talk about the issues of cross contamination that might occur when you do a gene test by mail - is the lab accredited? Or are they paying undergrads $2 an hour to do donkey work in the lab in their spare time? Mistakes happen all the time with this kind of testing.}
posted by JeanDupont at 2:26 PM on November 12, 2013 [6 favorites]

Oh, I think I got it - you think your cousin should have different Y chromosome marker to you, but you have the same one.

As I said, one reason this could happen is because the marker you share is common enough that it has come down into both sides of your family by chance - you don't have to be closely related for this to happen. Or there could have been a mistake in taking the sample, or anywhere in the testing pipeline.

Working out relationships within 3 generations requires looking at markers across multiple chromosomes, and doing statistical analysis that takes the frequency of those markers into consideration. There is no realistic way a test of one marker on the Y chromosome could tell you anything about shared ancestry within 3 generations. If that's what the testing company told you, that's very misleading.
posted by JeanDupont at 2:46 PM on November 12, 2013

Anecdote, slightly different than those previously shared:my grandmother's cousins, beautiful nurturing folks who were never able to have kids of their own, adopted a young girl named "Betty" in the 20s. For one reason or another, this adoption was more open than most adoptions of that generation and Betty later tracked down and had a friendly relationship with her birth mother. Who, as it turned out, was a close relative of my grandfather, who'd known and loved Betty for years at this point (my grandmother had been living with them when my grandparents began courting). So, for my mother and her siblings, this woman is a genetic cousin on both sides.

tl;Dr Crazy things happen, especially in small towns in the south.And of course, no genetic test can change who is a member of your family. There's much more to it than that.
posted by theweasel at 4:32 PM on November 12, 2013

Mod note: From the OP:

Jwells - Ancestry DNA Paternal Lineage test, 33 markers, my cousin and the other matched individual in their database shared a MRCA within 3 generations. All 33 markers matched.

muddgirl - Almost - cousin and I have the same grandfather, his father is my mother's brother.

Thank you all for your responses. I think what I will do is have a different cousin (one who has the Y chromosome that I am researching) but who does not share parents with MY cousin take the test. If it comes back the same way, logically, I can narrow the divergence to my great-grandparents or further back. If the other cousin has results which are more in line with my research, then I can assume that whatever happened happened at the level of my grandparents or his parents. In which case, I would definitely keep my mouth shut.

posted by jessamyn (staff) at 5:19 PM on November 12, 2013

This is my reading of the situation - though please, since I'm just someone curious about this sort of thing, no expert by any frame of reference, sum up my opinion as - Yeah, I think there's enough odd that it might be worth more research, testing, or better advice, though how many markers were tested is extremely important for accuracy. And I didn't see that number mentioned. If enough were tested, that would pretty much rule out it being random chance of being "someone with somewhat similar markers".

My understanding:
Anon and Cousin are supposed to share a grandfather - it's Anon's mother's father, but it must be Cousin's father's father. Anon's mother and Cousin's father are siblings. Also, if I understand correctly, Y-DNA passed down only through direct paternal lines, with very little change between generations.

I believe Anon is stating that the "perfect match", within three generations, is between himself (Anon) and Cousin, though they should be completely different, according to the history, as they have different fathers, and thus grandfathers and great-grandfathers.

It doesn't appear, from what Anon said, that presumed grandfather himself was tested - though if he's alive, asking him, their fathers, other male siblings, uncles, or cousins to be tested would probably be the way to proceed. Because the way I'm understanding this, either their fathers, their grandfathers, or their great-grandfathers are very likely the same person - and if not, it's likely to be just one more generation back.

Boils down to: there's some strong potential for unpleasant discoveries if enough markers were tested.
posted by stormyteal at 5:30 PM on November 12, 2013

my cousin and the other matched individual in their database shared a MRCA within 3 generations. All 33 markers matched.
...
cousin and I have the same grandfather, his father is my mother's brother.

I was under the impression that you were talking about a match between YOU and your cousin, not your cousin and someone else. Of course your cousin and you share grandparents on your maternal side - what I was implying is that you shared a grandparent on the paternal side. But it sounds like the match you're talking about is to someone else in the database, not to you, correct?
posted by muddgirl at 5:32 PM on November 12, 2013

Oops, OP and muddgirl posted while I was typing...

Ah - I didn't realize it was a match to some unknown in the database - I thought it was a match between you two. So... much of what I said is irrelevant.
posted by stormyteal at 5:35 PM on November 12, 2013

The reason that makes a difference, in my mind, is that you don't really have any idea of this other person's family history (even if you can access their genealogy research or something, you don't know the quality or the story behind the research), so jumping to rape or infidelity (vs. a name change) seems really, really premature.
posted by muddgirl at 6:03 PM on November 12, 2013

Oh my, without more evidence don't make assumptions that something happened and even if it did what is done is done. My grandfather had two kids before he married my grandmother, then took in 3 extra kids, gave them him his name and no one cared.
posted by OhSusannah at 6:43 PM on November 12, 2013

Restating the question so I am sure I understand: your cousin should have the Jones Y chromosome. His Y matches someone in their database who should not be on the Jones family tree based on your research.

So either his paternal parentage is not really in the Jones line, or there is an extra Jones out there that is unaccounted for. (Assuming the tests are accurate and precise.)

The likely explanation is that there is an extra Jones out there somewhere.

For example, my father had 4 brothers, and their father had one brother, and his father had one brother. I know that I have unaccounted for first cousins, I am fairly certain that I have unaccounted for second cousins, and given the profligacy of my Y chromosome, I probably have unaccounted for third cousins as well. Not to mention the potential progeny of those unknown individuals. IE, there could be many, many people out there who would be a Y match for me, but not show up on a family tree or even share a surname. All it takes is for great-grandpa Jones to slip one past the goalie in a random hookup in wartime France, Okinawa, or the ol' Roadhouse by the county line to end up with "extra" cousins.

And that doesn't account for the imprecision in the testing. 33 markers on a chromosome that contains far more information than that is a shot in the dark. It works good enough to rule out, but not so well to rule in.

There are also these probabilities, which seem much more likely than your 33 of 33 = 3 generations for sure analysis. http://www.familytreedna.com/faq-markers.aspx
posted by gjc at 4:37 AM on November 13, 2013 [2 favorites]

I should have told you a story: 1875 or so, my great grandfather Thomas Arrington (direct paternal descent) finally had enough of his dad and split. At the time it was legal to use either your father's OR mother's last name as your own surname, without any paperwork! And that's how my line of Wells' was created. The folks tracking all the lineages of Wells' genetically were literally stumped over us until we figured this out. We're Arringtons genetically, and it had nothing to do with an adoption or infidelity. Strange things happen more often than we think!

Of the Arringtons, we're actually off on our own as well. The other lineages don't match our own, as we seem to be the oddballs from Ireland rather than Scotland or England. And that has been the biggest eye opener. Some names are just so common that multiple independent groups ended up with the same surnames. Jones is fairly common, so I'd expect this to be true with that one in particular. 5.75% of Wales is a large group... 176,180 people. 0.75% in England (397,575 people) - and second to none other than Smith. 0.50% in the US (1,407,109 people)... which gives a total of 1,980,864 people. That's a large base. You could be dealing with a selection bias, as gjc indicates. Some lines know their history and don't get involved in genetic testing, others aren't as adept technology wise, and then there's the folks like us who do the testing for whatever reason.
posted by jwells at 5:52 PM on November 17, 2013