View Full Version : AncestryDNA Tree Experiment

11-14-2016, 08:15 AM
Some time ago a possible "2nd to 3rd" cousin named Erin invited me to share genomes at 23andMe. It turned out that we shared 2.02% across 9 segments. Because both Erin's father and mine had also taken the test, we were able to determine that our connection was on my mother's side and on Erin's father's side.

Erin's father's first name was Rodney, but it was his surname that caught my attention. It was the same as that of a man who married one of my grandfather's sisters. I realized that if Rodney happened to be the grandson of this sister, then we would indeed be 2nd cousins -- and Erin would be my 2nd cousin once removed.

Unfortunately, it appeared that this was not the case. There was a connection that involved this couple, but it was by only by marriage. Rodney was the grandson of the brother of the man my grandfather's sister married.

Naturally, it occurred to me that this could have been a case in which the child was raised by a different family member, and not told. But all indications seemed to be that our "connected couple" never had any children at all.

Nevertheless, the marriage was tangible evidence -- outside of the DNA -- to show that our families were connected in some way. But I figured that if the answer was not with this couple, then likely there was some other sort of NPE.

This seemed even more likely after an additional set of discoveries made me wonder if such an NPE might have been on my side and not Rodney's. I was showing dozens of Sizemore, Bowling, and Asher cousins for whom I could not account. This was in spite of the fact that I "knew" most of my grandfather's ancestry back to his great grandparents.

Rodney's grandfather, on the other hand, had a Sizemore grandmother and a Bowling great-grandmother. If, somehow, my grandfather were actually a brother to Rodney's grandfather, it would explain both why we show up as 2nd cousins and why I have these unexpected Sizemores, etc.

Of course, I didn't have any explanation for why my grandfather wouldn't have been raised with Rodney's grandfather, if they were brothers. I considered the possibility of some sort of affair, but it made no sense in either direction.

On the one hand, an affair between Rodney's grandfather's mother and my great grandfather wouldn't explain why my grandfather would be raised by my great grandfather and his wife.

But on the other hand, an affair between Rodney's grandfather's father and my great grandmother wouldn't account for the Sizemore connections.

At this point, let me mention that there has always been one "mystery" surrounding my grandfather's birth. Reportedly, it took place in Miami County, Ohio, in 1904. That's the are my great grandparents were from.

However, it wasn't where they were living at the time. They were in the deep south, and had to travel back to Ohio at a time when my great grandmother would have been quite pregnant. Again, this was in 1904. Perhaps they wanted to be with family?

However, very recently I learned that no record of my grandfather's birth in Ohio was made at the time. In fact, it was not until 1963 -- some 59 years after the fact -- that my grandfather himself caused his birth to be recorded. Apparently, he needed a birth certificate for something and had to make a request to the probate court of Miami County. He did not appear in person, but sent a notarized application from his home in Biloxi, Mississippi.

The application asked for the names of any attending physician or midwife, but my grandfather's answer was "unknown". His answer to the question of how many other living children his mother had was "two", but it does not seem that he was required to provide their addresses for corroboration. (In any case, they would have themselves been fairly young at the time.)

Okay, so far I've focused on what I learned through 23andMe, through communications with Erin and Rodney, and through some internet searches. I haven't yet spoken about Ancestry and my experiment, but I'm about to.

Very recently, Ancestry showed me 2nd cousin match, a man named Bob. Bob and I shared "265 centimorgans ... across 15 DNA segments". My only higher matches were to a known 1st cousin once removed ("491 centimorgans shared across 26 DNA segments"), and to my daughter ("3,453 centimorgans shared across 54 DNA segments"*).

But the best part came after I made contact with the cousin -- or more precisely, with his wife, who is the one who manages the account. Rodney's grandparents -- the ones who may also be related to me -- are also Bob's grandparents! That is, Rodney and Bob are 1st cousins.

So this brings me to the experiment. Because I'd taken both Ancestry's older autosomal DNA test and its new one, I have two "versions" of myself at ancestry. I initially had both of them tied to my name in one tree, but for the experiment I created a new tree and my more recent DNA test to myself in that tree. (Leaving my original DNA test connected to myself in the original tree.)

So what have I seen so far? Well, the new tree doesn't have any associated DNA Circles yet. My original tree has twelve, perhaps nine of which are through my maternal grandmother. (The other three are through my father.)

However, for the test connected to the new tree, I'm now showing some 85 tree hints. I have 99 hints for the original tree, the vast majority of which are on my maternal grandmother's side. But because neither my grandfather nor my maternal grandmother appears on the new tree, all of the hints involve my maternal grandfather ... and almost all of them involve Sizemores or Bowlings.

Of course, this isn't surprising in one way, because of who I'm showing as my grandfather's parents. But remember, the hints are connected to people with whom I do share DNA, and who are surely related to me in some way.

This doesn't include Bob, since his tree is private; though it would if the tree were public. It does include a woman who shares "58 centimorgans ... across 5 DNA segments" with me.

The tree hint suggests that she and I are 2nd cousins once removed. For that I'd expect about twice as much DNA. However, I believe there may be an error in her tree. It appears that an uncle and his nephew both had the same name, and my cousin shows herself as descended from the uncle. But she lists the year of birth that would apply to the nephew. If she got these two confused, she's still descended from the same common ancestors, but is a 3rd cousin to me instead of 2nd cousin once removed.

I still can't explain why my grandfather might given away. Both the ones who might be his parents, and those who actually raised him, had other children. (In fact, his possible birth family shows another child with the same birth year, which had me thinking about "twins separated at birth", or maybe some sort of switch. But in 1904, neither would have been born in a hospital.

Additionally, why go to the extent of either making the train trip, or concocting a story about such a trip? Especially when there were other children who could potentially "spill the beans"?

*EDIT: According to Roberta Estes, Ancestry's software is set to reject some potential matches because they're considered "suspect" -- too "matchy". This may be happening with my daughter. As her father, I really should show just 23 shared segments. (Or 22, if they don't include the X.) Instead, they're reporting 54! (If I remember correctly, 23andMe shows 24 or 25. That's still wrong, but a little less crazy.)