Sunday, October 14, 2012

Genes Day - A long overdue DNA update

I am finally finished updating my database with all the Groff family files I've collected (and some I dug up while going through those I've already gathered) so it's time to get back into DNA research! There have been some big developments over the last few months.

I realize I left on a cliffhanger (well, hillhanger more like). I did get the results in for that Harris cousin, and she turned out to not be related to us after all. Bit of a letdown, as these tests are not cheap, but it is what it is. I'm also still trying to figure out the ethnicity calculations for my grandma's test, as everywhere I post the results, I seem to get different answers - 23andMe, GedMatch, Family Tree DNA all have different answers for the same data. There's a lot of factors that play in to the calculations, I get that, and I don't know that they'll ever be able to say "this part of your DNA is German, this is Irish, etc" with any degree of certainty. But there should be at least a general level of consensus shouldn't there? How do I know who is right, when all the answers conflict so much?

On a related note, I got a message from a new match on my mom's DNA test, wanting to compare genomes. At first, I figured this would be like all the other matches I get emailed about - one tiny segment barely large enough to qualify for a place at the end of mom's matches list. Instead, I was shocked to see that this was mom's third closest match, second only to her mom and first cousin! The new match shares a total of nine segments, for a total percentage of 2.15%! That puts the match right between the expected percentages for a second cousin (3.125%) and a second cousin once removed (1.563%). I wonder how you could fall in between those numbers like that? Anyway, I compared her data to my maternal grandma, and there was no match at all, so that means this is a match on my maternal grandfather's side. I got even more excited when I saw that one of the match's ancestral surnames was Hammer - the maiden name of my mom's paternal grandmother! So I hurriedly wrote this match an email with some info on my Hammers and...nothing. No response. After three days, I wrote a follow-up email, and this time...I got nothing. Again. I can't do much else with this connection until she writes me back. So, with the excitement slowly fading into "why the heck would you want to share genome data but not email me?" I've decided to pass the wait by looking back into some of my other tests, and going deeper into those results.

My wife Lisa also took a DNA test recently! I FINALLY got my AncestryDNA invitation and decided it would be best spent by testing a whole new branch of my family tree - hers. I'm really glad I did, because it's given me a chance to try out a new DNA product on a family line that I haven't done any DNA research on yet. Her results have been interesting, to say the least.

First up was her ethnicity. From what I knew of her ancestral background, I expected 88% European and 12% Native American, as she has one great-grandparent who was full-blood Native American (with a registered tribal membership and everything). What I got was 73% Scandinavian, 15% Persian/Turkish/Caucasus, and 12% uncertain. I was completely puzzled - from I've seen of her ancestry so far, she doesn't have any Scandinavian ancestry. I've read about Ancestry showing those with UK roots to have a higher percentage of Scandinavian heritage than may be completely accurate. But 73%??? With not a single ancestor or name indicating anyone from any Scandinavian country? Even if somehow every English ancestor of hers was a direct descendant of Vikings, this would seem too high. To be perfectly honest, I was very disappointed when I read that, because it made me think "well, this is almost completely off. Now what do I do with it?" I know Ancestry has some upgrades in the works, so I'll reserve final judgment on their ethnicity analysis for now.

The Persian/Turkish/Caucasus part of her ethnicity, I have no idea where that could be from either. It may be there's a break in the chain in one of her lines that no one knows about. That's always a possibility, I understand that. One of her 3rd-great-grandfathers was illegitimate, with no known record of who the father is, but that wouldn't account for 15% of her genetic heritage even if the father was a full-blood Turk. As far as I know, all of her ancestors (aside from the Native Americans) were from northern and central Europe - predominantly England and Germany, probably Ireland as well. So there's another big chunk of her ethnicity that just doesn't make sense, unless we have her entire tree wrong. And I hope and pray that that's NOT the case.

As for the uncertain, I'm willing to bet that that's her Native American side. The percentage is just right, and given North American tribes' reluctance and even refusal to do DNA testing, it makes sense that this part of ancestry would come up as "uncertain."

One of the really cool things with AncestryDNA is it compares your trees with those of your matches, and can tell you surnames you have in common. If it can pinpoint an ancestor in common, it will tell you who they are, and show you the lines of descent to each of you, and what your resulting relationship is. You can then compare that prediction against the prediction based on the percentage of shared DNA. How cool is that? Sadly they won't tell you how much DNA you share, though Ancestry has stated they will be releasing the raw DNA data to its customers sometime next year. Hopefully they'll do it in a way that will allow the data to be used in GedMatch, FTDNA, and other places. So far, I've found two matches that have an ancestor in common! One of them traces back to Lisa's 4th-great-grandfather on her father's side, William Qualls. The other is a closer match, tracing back to Lisa's 2nd-great-grandfather on her mom's side, Thomas Berry Hudson. The Qualls relative would be Lisa's fourth cousin once removed, and the predicted relationship is 4th-6th cousin, so that seems right on the money. The other one would be a second cousin once removed, but the predicted relationship is fourth cousin. A second cousin once removed would normally share about 1.563% of their DNA, while a fourth cousin would share about .195%. That's a pretty big difference, one where having access to the raw data would really come in handy. I need to set aside some time to contact these persons and see if we can't compare notes and figure out exactly how we connect, and what we each have in terms of documentation.

While looking at all of this, I realized something. I've yet to fully document everything in my wife's tree as I received it from her family when we got married (as you know, I've been working the last 18 months just to document my own tree with the stuff I've already collected and I'm still not quite done with that project). But I realized her tree is really different than mine in one major way - I've documented her 16 great-great-grandparents, and every one of them was born in the US. Some of the links to her 3rd-great-grandparents are a little shaky still, but all of 24 that I have documented so far were born in the US as well. Contrast that to my tree, where three of my 2nd-great-grandparents were born in Canada (two of which were children of European immigrants, with the other being a grandchild of immigrants), four were born in Europe, and another three were born in the US but were children of European immigrants. That means that while all of my wife's lines have been in the US since at least her 2nd- or even 3rd-great-grandparents' time, 10 of my 16 lines in the same generation had been in the US only one generation or less. It's fascinating to think of the complex choreography of ancestors making their way to the US at different times, over different routes, probably for vastly different reasons. I am grateful for what DNA and genetic genealogy has taught me about my ancestry so far, and am greatly anticipating what insights and discoveries await!

1 comment:

Gibson Family said...

Ok, so the Harris' are probably not be related to us, surprises with aunts and uncles and grandpa’s aren’t really our grandpas, then we discover Lisa is Scandinavian, Persian / Turkish / Caucasus. (Wow - that's quite a little melting pot, huh!) I thought she was Irish! Was she surprised?

I dunno about this DNA stuff - "seems like magic and it scares me", to quote Jade. It sure rattles some cages and can disillusion you. I think I like just thinking I know who I'm related to and gonna forget the truth. Truth, it's just plain over rated!

But thanks for all of your hard work, Brandt. You are amazing and digging up dirt on all of us! Sure glad I don’t have any skeletons in my closet, or do I????? Love you!