A closer look at AncestryDNA
The exhibition hall at any genealogical conference is always a fun place to be and to hang out. So many goodies — new books, new products, new services — you almost don’t know where to look first. One thing I knew I wanted to check out at this year’s NGS conference in Cincinnati was the Ancestry.com booth and the new autosomal DNA test offering, AncestryDNA.
Fortunately, I was able to sit down for a time with John Pereira, vice president of DNA at Ancestry.com who runs the DNA Product & Marketing team. And, yes, like a bunch of other folks who are intensely interested in and writing about DNA, I was handed a free DNA test kit (and yes, I’m gonna do it, since I have yet to meet a DNA test I wouldn’t take, and no, Ancestry isn’t getting a pass on any tough questions as a result. I just believe in full disclosure).
So… with that out of the way… here are some things I was able to clarify and some things that have yet to be decided or clarified about this new test.
Right now, the autosomal test is only being offered to Ancestry,com subscribers at an introductory price of $99.00. Not everybody who wants to be tested right now is being tested right now (see availability, below), but Pereira said all Ancestry subscribers who get their names on the wait list during the introductory offering will get the introductory price.
Where will the pricing go after that? “It hasn’t been decided yet.” When will long-term pricing be announced? “We just don’t know yet.” Will it stay at $99 for Ancestry.com subscribers? “This is an introductory price.”
Again, the test right now is only being offered to Ancestry.com subscribers. It won’t be offered widely to non-subscribers until the demand from Ancestry.com subscribers is met. The lab facilities can only handle so many tests at a time, so even subscribers are being put onto a wait list right now.
Will non-subscribers be able to test with AncestryDNA? “Yes.” When? “We’ll know that when we see how the demand from subscribers gets met. … Our design is to take care of our subscribers first.”
Subscription or not
“No subscription will ever be necessary to access results.”
The promise is that test results and match lists will always be accessible whether the test taker is an Ancestry subscriber or not.
But some of the AncestryDNA results system is so tightly integrated with subscription-based Ancestry services that it’s clear non-subscribers will see a more limited set of functions and features (and that non-subscribers will be encouraged, and not at all subtly, to subscribe).
“It’s just that some of the information we’re providing is information our subscribers pay for. They’re entitled to access it all, and those who don’t subscribe won’t get the information that’s subscriber-only. But the results themselves will always be available.”
“We understand it’s important to serious genetic genealogists and we’ve got that under review. We’re working to figure it out.”
Pereira emphasized the desire to make DNA results as easy as possible for all genealogists to understand: “We’re trying to present the DNA information in a form that’s understandable and simple for everyone.” The game plan for the product: “simple, actionable, integrated.”
But, he said, that priority didn’t preclude providing more down the road. “This is a work in progress.”
Data from other test labs
“Yes, absolutely, we’re considering how we could allow people who’ve tested with other companies to transfer their results to the AncestryDNA database.” Will it happen? “We sure hope so.” When? “We couldn’t even take a guess.”
Acquisition of GeneTree
Data acquired from GeneTree and the Sorenson Molecular Genealogy Foundation included “an unparalleled sample set” with not merely DNA samples but samples validated by pedigrees and paper trail genealogies. The data greatly enriches the ethnicity and admixture information pool available to AncestryDNA. “We expect to provide much more detailed admixture results because of this acquisition.”
Privacy agreements entered into between persons tested and Sorenson or GeneTree are legally binding on Ancestry as the acquiring company. This doesn’t answer the issue of folks feeling put out that they tested with a non-profit research foundation and ended up with their data being owned by a commercial corporation but it’s all the law requires.
There is no current plan to integrate GeneTree test results into the AncestryDNA system. The GeneTree results system will remain up and running “for now.” Those whose test kits had arrived at GeneTree by the acquisition date will be reported at GeneTree. Those whose test kits had not arrived will receive refunds.
“We want to make GeneTree customers happy to the extent that we can.”
Integration with Ancestry user family trees
Ancestry is “keenly aware” that integrating DNA results with undocumented user-supplied family trees poses potential problems. “It’s clearly up to us to make sure people understand the tree matches are just a tool, just clues for people to use to make their own decisions,” Pereira said. “We have a big educational task in front of us. We have to make sure people know that it’s not just DNA. It’s DNA and records and research and more.”
“We hope we’ll get help and input from genealogists on how to validate results, how to show users that DNA is just another tool and can’t replace good research.”
Family tree matches will be presented the same way for DNA results as they are for other searches: if the tree is private, a match will be able to send a request to the tree owner for contact and the tree owner will be able to decide, globally, or person by person, to allow or disallow access. Similarly, contacting matches can only be done through the internal Ancestry system; email addresses will not be disclosed by Ancestry.
“It’s a global rules set we’re working under. All the same privacy settings will carry over no matter what part of the Ancestry system a person is seeing.”
Where to next?
So… how about full genomic sequencing? Is Ancestry looking in that direction? “We’re going to be looking hard at every new development especially since it is just crazy and exciting how costs are plummeting.” But, Pereira cautions, “we have to consider just what it turns out we can learn from full genomic sequencing that’s really valuable for our genealogy customers. Does this make sense — will it provide data that, for the typical user, will really advance genealogy?”
My own take
I remain deeply concerned that integrating DNA results with totally undocumented unvalidated user-supplied family trees is going to end up with an awful lot of people absolutely convinced of facts that just ain’t so.
Don’t get me wrong. I love the theory. I wrote about the potential for just this kind of link-up last year for the National Genealogical Society Magazine1 after talking to Blaine Bettinger of The Genetic Genealogist. If it were ever executed well, linking family trees and DNA results would be the best thing for genealogy since sliced bread.
My problem is that I’m not convinced that the AncestryDNA system can do it well simply because of the vast numbers of demonstrably inaccurate user-submitted trees that are already in the Ancestry system. Telling two people who are DNA matches that their family trees match on an individual who really isn’t an ancestor to one (or even either) of those people isn’t going to move the ball forward so much as an inch.
Take my own Baker line for example. There are still dozens of user-submitted trees on Ancestry.com that say that Thomas Baker (c1711-c1777) of the Pamunkey Neck area of Virginia descended from Alexander Baker who came to Boston in 1635. The whole “we descend from Alexander Baker” theory has been totally disproved by years of YDNA testing (not to mention good genealogical research).
But when one of these “yes we do descend from Alexander” descendants of Thomas Baker matches any Alexander Baker descendant (and they will, not because of the Baker ancestor but because of some other ancestor they share that they don’t know about), what I’m afraid of is that they’re going to think they’ve proved the relationship — to Alexander as well as to each other — when the autosomal test is doing and can do no such thing.
How are we as genealogists going to answer “but AncestryDNA says it’s so”?
I don’t know. And it worries me.
That being said, I wouldn’t pass up the chance to detect any autosomal match in my own family and AncestryDNA adds to my chances. But when I butt heads with one of those Baker descendants in the future, I’m going to ask John Pereira to help me explain that no, AncestryDNA really didn’t say it was so at all.
- Judy G. Russell, “Autosomal DNA testing,” National Genealogical Society Magazine, October-December 2011, 38-43. ↩
Your point about the challenges in error-prone user-submitted trees on Ancestry regarding matching on the Ancestry-DNA testing … that was mine too, and I’m supposing that this is being brought up to John Pereira and/or others in the loop there. Thanks for a very relevant post!
There’s no doubt that the Ancestry folks are aware of this. I don’t know how it can be anything other than problematic given the … ummm … bounty of inaccurate trees.
Thanks for providing this update. Did Ancestry indicate if they would be maintaining the SMGF website?
It seems to me that Ancestry have breached the terms of the SMGF consent form by using the data from SMGF participants without obtaining their consent first:
Those who submitted their DNA to SMGF did so on the express understanding that they were contributing to a research study conducted by a “not-for-profit organization”. The consent form explicitly states: “You will be told about any new information that might change your decision to be in this study”. It would appear that no one who has submitted a sample to SMGF has been contacted by Ancestry to request their consent to have their data included in a commercial database.
For anyone who has submitted a sample from Europe, Ancestry are also in breach of European data protection laws which require that “data should only be used for the purpose stated and not for any other purposes”:
Whether or not European law can be applied in America, it all seems highly unethical and a complete betrayal of James Sorenson’s vision.
I regret that I didn’t ask specifically about the Sorenson website. My impression was that the website and its assets were included in the acquisition but that’s not from anything specific that was said.
Great article Judy, as always. I’m confident that eventually Ancestry will return raw data to participants, perhaps for an additional fee(?). As for the error-prone trees, I’m convinced that DNA will help identify and correct these errors as more people collaborate and the tools for comparison and sharing improve. In terms of genealogy, none of us are “DNA islands” – our ancestors’ DNA has traveled far and wide and is infinitely more powerful than the mistakes we’ve made in our paper trails.
I’m also confident about eventual raw data access. But not as confident as you are about the issue with the trees. Example: how to convince a genetic cousin that though I have person A in my tree and he has person A in his tree, he really doesn’t descend from person A at all and our common ancestor is in an as-yet unidentified line.
Although we could probably talk for hours about this, I think this would be my 1 minute reply: Why do you have to convince him? His tree is available for you to view, and you can investigate any of the paper trail yourself. Besides, you and genetic cousin C (and D, and E, and…) will sleuth out the issue together independently of the knucklehead. Further, even if every single tree in Ancestry.com was perfectly based on the best paper trail, we would still have these types of errors simply because even the best paper trails are often based on lies and intentional omissions that our ancestors wanted to keep hidden in their lifetimes.
I understand, Blaine. It’s the sheer number of perpetuated errors that can be daunting, along with the impact of what will look to many like scientific confirmation. When science appears to put its imprimatur on bad genealogy, there may be even more folks who say that the good paper trail must be wrong.
My test kit was waiting for me when I returned home from Cincinnati. It’s now in the mail. From the success I’ve had with GEDmatch I have high hopes for this, though as Judy pointed out just because a match and I both have “person A” in our trees that doesn’t mean that is where our shared DNA comes from. What it will do is give me a point to focus on to see how well our paper trails hold up.
Yup, that’s the benefit. And I have high hopes for folks like you who’ll use the info correctly.
I am baffled why they are making it so difficult for GeneTree Y-DNA customers to integrate into the Ancestry system. The markers for the 46 marker test are identical. It shouldn’t take rocket science to do it. I am also upset that one of my tests en route from the UK on the day of acquisition is being rejected rather than processed. It is difficult enough to find people to test, but to return to the same person and ask for a retest is ridiculous. Ancestry should be able to process these kits with no problem.
I wish I’d asked that question, Jim!
I am excited about my test, (just sent in my sample today) and believe like with all things unknown, you have to start somewhere and also “production must be profitable”. I am an “Amateur Genealogist” and have been for 20 years. I recognize the inconsistencies on Ancestry as I continue to find mistakes everywhere and even in my own tree. I submit that it is value added to know of the errors or to have them out in the open and in using deductive reasoning you know almost immediately that something is aloof. I think the future will triangulate with current unknown variables and give forth a solution to this problem if not at the least provide a framework for Genealogists to work in on solving such bogus data entry. Empirically I see on some of my lines the same honest mistakes being made(myself included)by us amateurs but through trial and error we will get to the right answer eventually or the real dead end. In the long run America and the world will realize a widened sympathy towards one another as we rediscover our common goals and in that is the true beauty of the reality it provides. Thanks for the opportunity to state my opinion.
I hope you get all you want from this test, Thomas, and I am delighted by your attitude. If everyone who had a family tree online was as committed as you are to making sure it is accurate in the long run, then a great number of my fears about bad data would be gone.
John Pereira’s statement in regard to the necessity of a subscription could be interpreted as a bit misleading: “No subscription will ever be necessary to access results.”
I am aware that he clarified his statement a bit more, but when he references “results”, I guess he just means the “Genetic Ethnicity” pie chart since the list of matches is useless without a subscription. A person cannot view their matches’ trees or contact them. At the moment, it appears participants without a subscription are being offered a “complimentary 6 month membership to Ancestry Connections” to allow full access to their results. If no subscription will be necessary, what is “Ancestry Connections”? I wonder what happens after the six months is up?
Thanks for the great article and additional information on AncestryDNA’s product.
Great questions. Let’s see if we can get answers!
Received results. 1) Absolutely no credibility if they can’t provide the science & raw data behind their conclusions. 2) Under “birth locations in your tree” they seem to pull a few random people off the tree and plop them on a map–of what use is this? Think it will lead the uninformed to assume those particular ancestors provided their DNA. 3) “simple, actionable, integrated” –It is the job of AncestryDNA to educate the public about the science not assume a paternalistic approach & create a stupid marketing meme. 4) Of the 4 matches for 4th cousin, I shared a 6th great grandmother with one person and nothing from the others. Again, zero information on how they connected these people at a 95-96% confidence to me. Conclusion: Extremely disappointing and a waste of time. Am considering purchasing a test from Family Tree DNA.
Before you decide on more testing, make sure you carefully review what data you’ll receive and in what form. You may find one testing company is better than another depending on your personal preferences as well as on the science.
I think ancestryDNA is NEVER going to allow people to download the DNA raw data. (at least not for free)
They are counting on people paying $99 upfront and then most likely charging a monthly subscription fee for “Ancestral Connections” to receive the matches.
In no time you will have spent more money with ancestryDNA than you would have with the one time fee charged by FTDNA and 23andMe.
In addition, FTDNA and “23andMe” both give you the raw data.
You can then use gedmatch and many other tools to do as you wish with your DNA results.
Some have even confirmed their FTDNA results with “23andMe” and vice versa.
This is not even possible with ancestryDNA.
Do you remember the stock picks made by a monkey throwing darts at board?
Someone should ask Pereira if he thinks this would work for DNA matches too.
I suspect most AncestryDNA customers are already Ancestry.com customers, so the subscription fee is not an issue. But there’s no doubt that getting the raw data to work with is key to serious genetic genealogy and only time will tell whether Ancestry will commit to the serious genetic genealogist or not.
The autsomal DNA test through Ancesty.com is useless. Don’t waste your money. If you are a genealogist/family researcher who is just starting out in your research it may be somewhat helpful but for someone who has been doing research for over thirty years it is absolutely useless. It will give you many trees of people who are just starting their research which is of no use to you. It’s simply a money making project for Ancestry to gain new members.
I can’t agree that it’s useless, Teresa — any chance at a match has its value. But at this stage, until more work is done and the raw data is available, it wouldn’t be my first choice, no.
I absolutely agree with the “person A” logic.
Every one of my lines for generations lived, married, and died in southeastern Pennsylvania. Six of my 4th cousin” matches are connecting me to them via a woman I’m related to through adoption. Despite the fact that we could be connected thorough any other of hundreds of individuals, these individuals don’t want to believe that she isn’t our common ancestor. Ugh.
That’s my big concern, Heather, and has been since Day 1. I’m very much afraid that the tree approach is going to have a natural tendency to cement bad genealogy (“but AncestryDNA says it’s so”) rather than encourage good genealogy.
I have now purchased 5 of the $99 kits. I have my results, am waiting on my mother’s, and still need to obtain samples from an uncle on my father’s side, a more distant relative descending separately from a “mystery branch” also on my father’s side, and my wife. I’m hoping that by getting separate maternal and paternal family results, I can divide the unknowns and speculations into tighter groups and maybe make some headway.
To maximize this opportunity, I expect that Ancestry must either release the results and significantly improve the match comparison tools. I noted your concerns that the DNA results may entrench false beliefs based on erroneous tree data. Without much sharper DNA analytic tools, I agree. However, with tools that allow comparison of the actual matching segments among purported matches, it should increasingly sharpen the DNA connections and improve the mapping to family tree entries. I expect that Ancestry should eventually be able to inform respecting the reliability of family tree entries.
This is very interesting stuff and I thank you for reporting on it.
Scott Lee Johnston
If they develop the tools and/or make the raw data available for third party tools, then I can at least see hope that you’ll be right, Scott. The fact is, a combination of DNA and solid paper trail genealogy is necessary — and it’d be awfully nice to have the tools to do that all at once.
I share your concerns about Ancestry’s use of their Public Member Trees as a source of information for a participant in their Autosomal DNA program. Given the number of errors I have seen in those trees I have little trust in their validity. In my Payne family I would say that over half the Ancestry trees contain serious errors—children born before marriage date and in different location them stated for parents, children born outside of mother’s child-bearing age, wives listed that records show married another, sons listed that Y-DNA has proven not related, etc. All are errors that a little basic thought and research should have prevented. When I contacted Ancestry some time ago with specifics on some of these errors I was told that Ancestry was just a distributor of the information in the Public Member Trees and, thus, not responsible for the content.
I was told I could contact the tree owner through email or make comments for individuals where I thought there was an error as a way to make corrections. I have sent hundreds of emails and made many, many more comments on tree entries for errors that I can document. Out of each hundred I have contacted about two or three respond and, as a guess, four or five make changes.
I have suggested to Ancestry there should be means to “red flag” entries that are incorrect or to use “Post-um’s” as their Roots Web World Connect site offers for reader input. I stated as a bare minimum each entry should be dated, again as the World Connect sites are, so readers would know how up-to-date the material might be. I was told my suggestions would be passed along. I have not heard back so I assume Ancestry does not see my concerns as a problem.
Now, with Ancestry providing Public Member Tree information as source material for an individual’s family history, it looks to me like it is a different ball game. I am not an attorney, but it seems to me that if they are referring atDNA test participants to a person in a tree and saying this is your cousin and, by implication, their family tree gives the participant a view of the participant’s family history, then Ancestry has an obligation to ensure a degree of accuracy for that tree. I expressed these concerns in an email to Ancestry a while back, but, of course, did not receive a response. I also said I was thinking about having an atDNA test just to get my name on the list for a possible class-action lawsuit in the future.
Perhaps what needs to be done is to set up a web site that names those listed in Ancestry Public Trees with errors on their family history so test participants could determine if the information they received from Ancestry was correct or not. Perhaps it real value might be to highlight the inaccuracy of the information on Ancestry Public Trees. I get twinges in my stomach each time I see the Ancestry TV advertising that says go to the Ancestry web site and instantly learn all about your family history.
As I see it the push for Autosomal DNA testing is going to compound the problem of errors on Public Member Trees. Now people will say, “But my DNA tests prove I descend from person X.” as they use this information to prepare their own trees.
Many years ago a favorite computer expression was “Garbage In, Garbage Out”. It is just as true today. Unfortunately, in my opinion, much of the information on Ancestry Public Member Trees is garbage. So be aware that if you are thinking of having an Autosomal DNA test, you may be paying for garbage if your results are tied to Ancestry Public Member Trees.
I am pleased that you brought this subject up for discussion and to learn I was not the only one with concerns.
Your points are, all, only too well taken, Dick. Ancestry certainly doesn’t have any legal obligation here but its contribution to junk genealogy by tying the DNA results to notoriously inaccurate trees is not helping anybody here.