A relative just got results back from an Ancestry test. About what was expected but what I heard, second hand, is that she was asked what she knew of her ancestry and the test results pretty much agreed with that. The person who told me this was suspicious that the company is just feeding you back the answers you want to hear. I wonder whether she would have gotten the same results had she said she didn’t know anything about her ancestry.
Ancestry definitely doesn’t just feed you back what you tell them you know. But they are trying to keep people around to buy other services. They were, after all, a genealogy website long before they started doing DNA and they get a lot more out of people paying their exorbitant annual subscriptions fees than they do for a one time DNA-test purchase.
My result, and that of my father, grandmother and aunt all reflect the difficulty in telling regional Norwegian, Swedish and Finnish DNA apart and distinguish the region from the rest of Northern Europe. The matches confirm what I know, the “ethnicity estimates” which I already knew were more of a parlor game, wildly overestimate the Swedish component and adds in a huge chunk of British that doesn’t fit the tree at all.
The many people who get expected results just happen to not have any “out of ethnicity extramarital” ancestors, and their ancestors happen to belong to the groups that the machine learning algorithms are weighed towards.
You don’t have to tell them. A friend of mine who was sceptical took a 23andme test, didn’t tell them his background, which is mixed (different continents), and his name is, anything, misleading when it comes to his background. He doesn’t use Facebook or anything where some researcher would be able to look him up. The test came back completely accurately reflecting the main areas his ancestry is from. Just one case, of course.
It’s very likely we’re related, on your Dad’s side. My grandmother was really into geneology (even transcribed census and cemetary records), and so I have a lot of information about my ancestry, even before my parents did their DNA tests. I’m related, but not descended from, Levi Coffin, who ran one of the stations in the Underground Railroad.
My dad’s maternal grandparents came from Germany, and his father’s ancestors came from Switzerland (7 generations back), but his DNA test picked up mostly Scottish and English heritage.
I haven’t done a DNA test, and I would expect the results to be a mix from my parents.
Interestingly enough, my parents did theirs in 2012 and got an update last year. As more people get tested, there’s more information available. I guess there’s a distant cousin in Ireland, who we didn’t know about before.
My grandmother was never able to figure out exactly when her paternal line came over, because there’s a dead end, as the date and place of his birth is theoretically known, but there’s no information about his parents. So the date and place probably came from him, and may have been falsified.
As a biologist and a geneticist, I’m surprised that no one here is concerned about the possible legal ramifications involved with sharing your entire DNA profile with a private company, who then owns all your data and can do anything they like with it. The laws regulating this are extremely murky, with law enforcement already using it in creepy ways. Down the road it could lead to issues involved with insurance, health care, or genetic discrimination. Their terms of service can also change at any time, so whatever you agreed to when you signed up isn’t set in stone.
As a professional in this area, I would NEVER submit my DNA to one of these companies. At least not until the privacy laws regulating the field were significantly beefed up. And probably not even then.
(If I was desperate to have my genome sequenced I’d f*cking do it myself in my lab, we have a Nanopore sequencer that could be easily and cheaply used for the purpose).
As an aside, I’m amused by the number of people surprised by the amount of infidelity uncovered by these services. When I was in grad school I was involved with a genotyping project in southeast Asia, where for various reasons we were collecting genetic profiles of every inhabitant of the villages we were working in (way before genome sequencing, this was with microsatellite markers). The data were blinded, so while we knew the relationships between samples (husband, wife, son, daughter, brother, sister etc…) we didn’t know the identity of the people. There were many cases (about 5-10%) where I could tell just looking at the marker profile that the children in the house did not have the father they thought they did. The percentages we calculated (for our own datasets, we couldn’t publish the numbers) agreed with rates of infidelity across most human populations. It’s surprisingly common.
This is the reason why I don’t think much of the DAR and other such organizations. I’m happy that there are people are thrilled to be members, but I always wonder how many people are really entitled, that means via bloodlines, to be members.
The Mayflower Society in the U.S. is quite interested to get their hands on their members’ DNA profiles and expect that in the future they will grant membership based on DNA profiles alone. Sort of takes the fun out of the geneology part of it.
The Mayflower Society in the U.S. is quite interested to get their hands on their members’ DNA profiles and expect that in the future they will grant membership based on DNA profiles alone.
After so many generations, how strong could such a genetic link be? Or would it be another variant of “just one drop?”
Come to think of it, could such analysis turn up “wrong side of the blanket” shenanigans that far back?
Y-chromosome or mitochrondrial DNA matches would still be very strong; that’s how Richard III was identified, for example, although Bosworth Field was more than a century before Plymouth Rock.
If you are on paper a direct male-line descendant of So-and-so but have a different Y chromosome than other male-line descendants, then yep, there is something hinky somewhere (as showed up in the Beaufort line in the aforementioned Richard Iii case).
I just find it very unlikely the laws would change, or not catch up, as these companies start using my DNA for nefarious purposes, and not at the same time allow those nefarious actors to start testing the samples you give your doctor and/or let insurance companies require genetic testing.
The number of people who’ve done such testing is still a small minority, quite a large percentage of them are elderly, so the testing companies would burn their current business permanently for short term gain, and unless it was part of an overall trend towards letting them have all the samples they want, the nefarious actors would be burning themselves too, for a short term gain for a small segment of their business.
If you (generic you) have concerns like these, but still are interest in testing, I’d go with Ancestry. They’ve specifically chosen to focus their selection of variants to those without known correlation to biological outcomes. (These tests don’t comprise your whole DNA profile, just a small percentage of known variants.) They’ve also successfully fought law enforcement requests for their data. And they are more restrictive in what they tell you about your matches, to avoid the very small chance of you using it to deduce that your 3rd cousin is also a bearer of a particular gene.
I see during my drafting of this, @mozchron has ninja’ed me.
Remember the deal about Radio Shack selling off its data when it went bankrupt?
To each his own, but that’s why I would never get DNA testing…you don’t know where the data may end up and you’ll never get that toothpaste back in the tube. This one still blows my mind.
Good GOD this guy was prolific!
2 Crimes 2.1 Visalia Ransacker (1973–1976) 2.1.1 Burglaries 2.1.2 Shootings 2.2 East Area Rapist (1976–1979) 2.2.1 Rapes 2.2.2 Murders 2.3 Original Night Stalker (1979–1986) 2.3.1 1979 2.3.2 1980 2.3.3 1981 2.3.4 1986 3 Communications 3.1 Written 3.1.1 Excitement’s Crave poem 3.1.2 Homework pages and punishment map (December 9, 1978) 3.2 Phone calls 3.2.1 I’m the East Side Rapist (March 18, 1977) 3.2.2 You’re never gonna catch me (December 2, 1977) 3.2.3 Merry Christmas (December 9, 1977) 3.2.4 Watt Avenue (December 10, 1977) 3.2.5 Gonna kill you (January 2, 1978) 3.2.6 Counseling service (January 6, 1978) 3.2.7 Later calls (1982–1991) 3.2.8 Final call (2001)
That’s the guy that Patton Oswalt’s late wife wrote a book about. She died shortly before the dna uncovered his identity – I’ll Be Gone in the Night. HBO made a series about it. Horrifying interviews with some of the early survivors.
None of these sites check for ID. You can get results under a pseudonym.
I can choose to opt out but my mom and sister have used both of the popular sites so my cover is blown. I never committed a crime where my DNA would give me away but I was a sperm donor around 40 years ago. I was promised anonymity. Now they can find a genetic grandma and aunt. My only saving grace is that if they are anything like me, they won’t want to find me.
I sent a test off for my son, who has the same last name as me. No information whatsoever was given to the company. His test came back identifying not only that he was 50% Swedish (on his mother’s side, so no last name clue), but also nailed the specific region of Sweden where we know for a fact her relatives came from. It also specifically identified that he’s from early settlers to Upper Canada in the 1800’s which is (as far as I know from family records) very accurate.
Again, no information at all was shared with the company before results sent. No ancestry trees set up on their site, nothing.
I would imagine that they also are crowdsourcing some of the data they have in their database; I mean, if they have never gotten any data from Madeupistan, and you know you are from Madeupistan, then they can possibly fill in gaps in their database.
Nope. That is not how it works. The ethnicity estimates given are based on “comparison” to specific reference groups and each group has to be many individuals. I write “comparison” because, unlike with the cousin matches, your DNA is not compared directly to these individuals, instead the reference groups are used to run a machine learning algorithm trying to group all the individuals in the reference groups as best as it can based on the frequency of the various variants in that group, with the goal being to replicated the known groupings as best as possible. I don’t know if the exact details are published.
Your data is then run through the final algorithm chosen and it spits out an estimate of which groups you belong to.
They only update these algorithms every few years, so a new member from Madeupistan won’t change them. I’m not sure the algorithm will even tell the company that you are all that different from their current members. You will likely come up as belonging more to their “broadly Central Asian” group than any other group, but that’s also what they get from their carefully selected group from Alreadyregisteredstan, so it won’t be noticed.
Well, I look 100% European, and that’s what the results confirmed.
However, after I first got it, it said around 50% of my ancestry is from the UK area, and around 30% was German (with a smattering of other stuff). Since then, it’s switched completely. Now, 30% UK, 50% German.
Honestly, if you think you’re European, you probably are, and I don’t think your results are going to be terribly accurate in terms of specifics.
Then there was the health aspect, which also didn’t tell me anything I didn’t already know.
If I had to do it all over again, I probably would have pocketed my $100 and not bothered.
I would not recommend such a test to anyone looking for relevant health information or revelations in the ethnicity results. Personally I did it for the cousin matching. And although I so far haven’t gotten any information that expands my tree I did have a pretty complete one already, and it’s neat to see how much of it I now have DNA evidence for. Any extramarital twigs looks to be no closer than great-great-grandparents, and at that point it’s starting to become difficult to tease the truth out of the available data.
As I said earlier, my son’s results pinpointed a specific province in Southern Sweden, which happens to be exactly where his maternal grandparents come from. It also pinpointed the outer Hebrides in Scotland, specifically North Uist, which is where his paternal great grandmother is from.
I got pretty useful medical/genetic info years before I was able to get confirmatory info from a medical office, FWIW.