The AncestryDNA Witch Hunt

A word of advice: beware anyone who tells you to avoid AncestryDNA.AncestryDNA1

Many genetic genealogists, myself included, have had incredible success using AncestryDNA’s autosomal DNA test.  Personally, several of my own major DNA discoveries have occurred though the service.  Unfortunately, it has become popular among some genetic genealogists to deride AncestryDNA’s autosomal DNA test, and some recommend avoiding the service altogether.

While AncestryDNA certainly does have limitations, avoiding the service is missing out on a major opportunity and one of the largest autosomal DNA databases in the world.  This is especially true for adoptees; anyone that tells an adoptee not to test with AncestryDNA (or not to test with any one of the three major testing companies) should not be assisting adoptees.

Indeed, AncestryDNA arguably has the largest autosomal DNA database of solely genealogists.  Further, many of the people in AncestryDNA’s database are people who would not typically test at other companies, since Ancestry has so aggressively advertised the product to genealogists of all levels of experience.

Here are some of the major complaints about AncestryDNA, and my comment on those complaints:

There’s no Segment Data (Chromosome Browser). It’s unfortunately true that AncestryDNA does not share segment data, and that data should be available for those of us who would use it. Indeed, I want that data. Ancestry already has the necessary data, and it would take only a tiny bit of code to present it to users.  But the undeniable fact is that the vast majority of users would not – and perhaps more importantly could not – use segment data.

Those of us who are immersed in genetic genealogy on a daily basis often forget that most test-takers are novices, and would not use segment data or a chromosome browser even if it were offered. Indeed, most test-takers would have no idea what to do with segment data. (For example, I would love to see statistics on what % of customers at FTDNA actually use the chromosome browser or download segment data to any real extent – I guarantee it’s a very small percentage).  While this is not meant to be an excuse, it explains in part why AncestryDNA does not focus on a chromosome browser.

This may not be a very popular view, but I believe AncestryDNA’s alternative to segment data, triangulation, will be much more useful to the average user than segment data, especially in the long-term future (thinking years ahead when the database reaches critical mass – see my post “The Science Fiction Future of Genetic Genealogy” for a better picture). Many of the things that we use segment data for – chromosome mapping and confirming matches – will be done automatically for the novice using advanced triangulation.

But again, I am not saying I don’t want segment data, I in fact do. Professional genetic genealogists need this data for some of the work they do.

The Trees are Wrong (and thus the hints are wrong). Yes, of course some of the trees are wrong. We were all newbie genealogists once, and we all created terrible, undocumented trees (we just weren’t able to put them online).  And I have personally created “quickie trees,” trees for adoptees or other clients that I need for quick comparisons or for “fishing,” but that I don’t have time to completely document.

If you’re relying only on a tree you found online without doing your own research, then there are more serious flaws in your research than the DNA tree hints.

Indeed, even the most documented and sourced tree is likely wrong, although genealogists are loathe to admit this.  NPEs (non-paternal events) happen, and many were either never publicized or were lost through time. Further, there are simply errors in most family trees that result from inaccurate source documents, or other sources.  Ironically, it will eventually be the DNA that reveals the errors in our trees, even without documentation.

Of course, take all tree hints with a serious grain of salt, but don’t throw the baby out with the bath water. A professional genealogist knows to question all records, no matter how official or ordained, and tree hints are no different.

The Biogeographical Calculation is Wrong. This one frustrates me the most, and even experienced genetic genealogists fall prey to it. It’s not specific to AncestryDNA, but applies to all three companies.

AncestryDNAMany people complain because their biogeographical estimate – their “ethnicity” – is not what they “expected” based on, typically, their known genealogy.  However, except for the rarest of circumstances, there is no such thing as “expected ethnicity.” No one, not even the most experienced genealogist with the most documented family tree, can completely predict their genetic biogeographical estimate without some previous DNA testing (parents, self, other relatives).  (The only caveat here is the family that has roots in only one tiny biogeographical region, which will be exceedingly rare).

As an exercise many years before autosomal DNA testing was available, I performed my own biogeographical estimate using the known birthplace of each of my ancestors. I suspect many people do this, either on paper or mentally, and believe that the biogeographical estimate they receive from a testing company should bear some resemblance to this number.  Unfortunately, these people are wrong.

At each generation, only 50% of a person’s DNA is passed to their children.  That 50% is almost completely random. Further, at about 5-9 generations, ancestors start to completely fall off your Genetic Family Tree.  This means that your Genetic Family Tree is a small subset of your Genealogical Family Tree, and without DNA testing you have no idea what subset that is.  You cannot predict your biogeographical estimate. You can make a rough approximation, but you cannot reject a reasonable biogeographical estimate based solely on what you think it should be.

Now, for several reasons, please don’t leave a comment below saying, “well XXXX test found 5% Southeast Asian, and I have a documented family tree of my family in England for 1,200 years.”  First, no one has a completely documented tree, and no one knows every instance of our ancestors’ lives.  Second, I’m certainly not saying that any company’s biogeographical estimates are perfect, or even correct. There is incredible room for improvement.  What I am saying, however, is that you can’t discredit it because it’s not what you “expected.” That’s inaccurate, and very bad science. Discredit it for other, more accurate reasons: the parents and children don’t align properly, it’s based on too few samples to be accurate, etc.

Conclusion

So, in conclusion, beware anyone who tells you to avoid AncestryDNA for any of the reasons above. You’ll be missing out on one of the largest autosomal DNA databases, and all the discoveries that could come from that. I look forward to your comments below.

71 Responses

  1. Jason Lee 16 February 2014 / 5:01 am

    “But the undeniable fact is that the vast majority of users would not – and perhaps more importantly could not – use segment data.”

    Yes. But not everyone has to use segment data to benefit from segment data. Users frequently benefit from interaction those who have more experience and better skills.

    Moreover, just because 23andMe and FTDNA haven’t figured out how to make segment data accessible and useful to most users doesn’t mean Ancestry couldn’t. I would argue that with their trees, they would have a great advantage in this.

    Presenting segment data with something like a “Who else matches us here?” feature — along with the relevant trees — could make triangulation much more intuitive and useful for the average user.

    I hope Ancestry hasn’t decided to forgo the benefit of having thousands of users generating segment data insights that might be just beyond the reach of a computer algorithm. Don’t underestimate the customers!

    • Blaine Bettinger 16 February 2014 / 9:27 am

      I agree Jason, and that is my point about a triangulation tool, which the average user will find more beneficial than segment data.

      • Jason Lee 16 February 2014 / 10:32 pm

        I would hope that the above-average user would be a bit leery about using a triangulation tool that doesn’t provide any segment data. It would be a bit like grafting a branch from someone else’s family tree onto one’s own without checking for documentation. I would also hope that professional genealogists would advise caution in using such a tool. Perhaps Ancestry’s final product will assuage such concerns.

        • Rob Warthen 17 February 2014 / 1:20 pm

          I have to agree with Jason. There isn’t any “triangulation” since we don’t know how these other people match. The fact that two other people share the same ancestor doesn’t prove that they are your ancestor. If I understand the segment data, I can then help the next person understand and so it gets paid forward. Ultimately Ancestry does us all a disservice by not providing this feature, but their family tree matching service is what the other folks need to move towards.

          However, overall, fishing in 3 ponds is the best choice. The fact that many provide family trees is very helpful. I also 100% agree that anyone who says NOT to test there at all is doing a disservice. We in the adoptee community encourage the testing in all 3 major test sites. I don’t know anyone who discourages testing at Ancestry.

          • Blaine Bettinger 17 February 2014 / 1:30 pm

            Rob – Although no one has any specific information yet, the hope is that the new “Triangulation Tool” (or perhaps “Verification Tool”) involves segment data (even though it will probably be behind the scenes). In other words, the match will be “VERIFIED” in that it is a segment of DNA shared by 3 or more cousins descended from a common ancestor (similar to how you perform triangulation with segment data from 23andMe and FTDNA). This will effectively be mapping DNA to particular ancestors, which is part of their long-term goal. Again, keep in mind that I’m only speculating based on what I heard at RootsTech, so I could be way off base.

            However, if AncestryDNA is only basing “verification” on overlapping trees among people who share (any) DNA, then this is no different that what they are currently doing. So there has to be something different with the Verification or Triangulation tool, and everyone has speculated that it is segment data (although it will probably be behind the scenes).

          • Jason Lee 21 February 2014 / 4:36 am

            Using all three of the most popular DNA testing services is certainly optimal for those who can afford it. (Full disclosure: I buy kits from all three companies, including Ancestry when the price is right.) But some people are interested in testing quite a few of their family members. For genetic genealogists on a tight budget, Ancestry’s lack of transparency with segment data could be an important consideration in favor of testing at FTDNA and/or 23andMe.

        • Caroline 6 December 2017 / 10:24 pm

          My thoughts exactly you can’t argue that triangulation data is great for the average user if segmentation data is ultimately better. It is the job of ancestry.com and competitors to not only offer segmentation data but in a way that the average user can appreciate and understand. One truth.

    • Brett 10 October 2014 / 6:35 am

      This is from Ancestrys site and it appears that you can download the Raw Data. Am I missing something? “Raw DNA data is the data that’s generated in the lab from your DNA sample, before it is processed or “translated” into usable information for family history research. The AncestryDNA test produces over 700,000 pieces of data that we use to calculate your DNA results and match you to others. The data includes SNP (single-nucleotide polymorphism) locations and the genetic code found at SNPs in your DNA. Genetic information is encoded in your DNA as a mixture of four nucleotides indicated by the initials A, T, C, and G. Each SNP has a combination of two alleles that encode your DNA at that location and determine your unique attributes.
      The information that you’ll receive with your AncestryDNA raw data will include the ‘rs’ ID where possible, chromosome, and the base pair position on the human reference genome (GRCh37). The genotype (the observed alleles at each position) will be provided on the forward strand.
      The security risks to downloading your raw DNA data
      While we protect all of your information in our systems, once you receive the downloaded raw DNA data, that downloaded copy will not be protected by our security measures. There are several steps required before downloading your data to help ensure your data is protected. Downloading your DNA data will not affect your AncestryDNA ethnicity results, DNA matches or future updates nor does it include this information.
      An example of what raw DNA data looks like:
      rsID chromosome position allele1 allele2
      rs4477212 1 72017 A A
      rs3094315 1 742429 A A
      rs3131972 1 742584 G G
      rs12124819 1 766409 G G”

    • Sylvia 4 May 2015 / 11:26 pm

      I am one of those newbies. I have several road blocks and also some common new England family names,, like Holmes, I am female so my dna will not trace my paternal line and I am trying to find the correct Holmes. I have added different options to my tree and then tried to document them but they always fall short and I end up removing them. I hope my DNA test will help. Being a math person I break everything down on spread sheets. I want to be able to download my DNA Matches I would like to see all my DNA matches broken down by chromosome and matches, sorrt them by common denominator.
      Am I correct in thinking this triangulation tool works in this manner however using just myself and 2 matches? I can understand why most would prefer a program to calculate these things for them but that is like just importing someone else’s tree and not verifying the supporting documentation.

  2. Maurice Gleeson 16 February 2014 / 7:47 am

    Excellent post Blaine – I’ve tested with all 3 companies and to my mind AncestryDNA’s big advantage over the others is its easy access to your matches’ trees as well as documentary records to enable further traditional genealogical research. Only 20% of my matches at FTDNA have uploaded trees to the site (66% have uploaded surnames) and I imagine it would be much less with 23andme (it’s not possible to sort your matches by those that have a tree uploaded so I don’t know what percentage of my 23andme matches have uploaded trees). And at the end of the day, no matter what the DNA says, you will have to compare trees with your match, and if it isn’t there, that is a distinct limitation.

    However, there is one big drawback for me with AncestryDNA. It is still not available outside the US despite promises that it would be (living in London, I had to travel to New York to take the test) and therefore most of the database is US – few Europeans, few Africans, few Australians, few Asians. So it is an excellent resource for anyone with Colonial US ancestors but comparatively less useful for those trying to connect with ancestors who came to the US in the last 100-200 years, and less useful still for anyone without any US connections.

    It would be a step in the right direction if FTDNA could use their partnership with MyHeritage to copy AncestryDNA’s example and provide links to online trees for all one’s matches. And then, as you say, if they also developed a function that would allow automatic identification & processing of Common Matches (and matches with overlapping segments), and triangulation with online trees, that could give them the edge over AncestryDNA … possibly?

    • Helen 16 February 2014 / 9:20 am

      You didn’t really need to travel to New York, only a US postal intermediary (which I have done for both 23andMe and AncestryDNA, neither of which ship to my country).

    • Blaine Bettinger 16 February 2014 / 9:36 am

      Thank you Maurice. You couldn’t be more right – the real power of DNA is the combination of DNA and traditional research.

      You’re right, not being able to test outside the U.S. is a major limitation. I’m surprised that they haven’t offered it country-by-country as they clear any legal hurdles, which would make more sense than waiting until they could offer it everywhere.

  3. Debbie Kennett 16 February 2014 / 8:48 am

    What you say is fair enough for Americans but the biggest problem with AncestryDNA as far as I’m concerned is that their test is only sold in America. It’s fine for Americans who want to match other Americans. It’s of no use to anyone living in any other country in the world because they cannot buy the test even if they wanted to do so. I’m one of the few non-Americans in the Ancestry database as I managed to order a test in the early stages before they shut down the overseas orders, I don’t have a single leaf hint, and there’s nothing I can do with any of my matches. If Ancestry do eventually start selling their test in other countries people are going to be completely overwhelmed with all the American matches.

    • Blaine Bettinger 16 February 2014 / 9:54 am

      Debbie – I agree, that’s a major limitation that I hope gets remedied soon. In the long run, absolutely everyone would benefit from international testing.

  4. Diana Quinn 16 February 2014 / 8:54 am

    I am an experienced researcher; however, my DNA knowledge is weak and I do not know what to do with a chromosome browser, but want to learn. Jason Lee’s suggestion is brilliant – “Presenting segment data with something like a “Who else matches us here?” feature — along with the relevant trees — could make triangulation much more intuitive and useful for the average user.”

    I had myself and four family members tested at AncestryDNA and, between all of us, there are over 40,000 matches. My matches alone total today at 8,550. Out of all of those, I have only seven “tree hints.” One is dead wrong and the others are known family members. By searching for surnames and common localities, I have determined that I am researching in the right direction for two of my family lines, but have found no other useful information at AncestryDNA.

    Trees alone are not enough. I hope that AncestryDNA will continue to add tools that sort through all of its generated matches.

  5. Blaine Bettinger 16 February 2014 / 10:00 am

    Diana – keep your eye out for a triangulation/tree tool coming from AncestryDNA this year. They discussed it heavily last week at RootsTech. I’m pretty sure it won’t have segment data, but it should be immensely useful for genealogists.

    • Kelly Wheaton 16 February 2014 / 11:17 am

      The triangulation tool will be a welcome addition and I like everyone else hopes they surprise us with segment data although i don’t expect it. i recently sent out invites to my top 50 MATCHING tree HINT matches asking them to upload to GEDMATCH. One responded immediately that they already had and sent their KIT number. And the other two may. That one match allowed me to identify matching segments at FTDNA and 23andme. Since I literally have ten times as many identified matches at ANCESTRY than at the others knowing even a few more would be awesome.

      I do hope ANCESTRY goes international. I have been advocating ANCESTRY with caveats from the very beginning. One long ago post was about not throwing out the baby with the bathwater. I suspect if the triangulation tool doesn’t have an easy workaround to segment data ANCESTRY will hear about it and eventually they may give us what we want. Hope spring eternal!

      • Blaine Bettinger 16 February 2014 / 1:59 pm

        Kelly – sorry for using the “baby with the bathwater” language in reference to AncestryDNA, must be great minds think alike! 🙂 Congrats on the success with getting matches to upload to GEDmatch. It would be interesting to know how you asked them (I’m always looking for advice and recommendations for my “Begging for Spit” presentation!).

  6. Anthony Parker 16 February 2014 / 10:23 am

    I wish that Ancestry DNA would have an autosomal transfer like Family Tree DNA has so I could transfer my maternal grandparents’ 23andme results to them. We have done Gedmatch but not everyone has sent their raw data to Gedmatch (I have also found that Gedmatch’s admixture calculators,at least for us, are way off). I asked Ancestry about this,and they said they most likely never will have an autosomal transfer because according to the operator I spoke to, Ancestry looks at more SNPS than FTDNA and 23andme does.

    Ancestry could also have assistant collection kits for folks who are not able to spit easily.

  7. Anthony Parker 16 February 2014 / 10:25 am

    Wish that Ancestry DNA would have an autosomal transfer like Family Tree DNA has so I could transfer my maternal grandparents’ 23andme results to them. We have done Gedmatch but not everyone has sent their raw data to Gedmatch (I have also found that Gedmatch’s admixture calculators,at least for us, are way off). I asked Ancestry about this,and they said they most likely never will have an autosomal transfer because according to the operator I spoke to, Ancestry looks at more SNPS than FTDNA and 23andme does.

    Ancestry could also have assistant collection kits for folks who are not able to spit easily.

    • Angie Bush 16 February 2014 / 11:14 am

      I am with you, Anthony! In fact, I would rather have to ability to upload autosomal data from the other two companies to Ancestry than a chromosome browser or segment data. There are individuals in my family who I tested with FTDNA prior to their passing away, and I think having them at Ancesty would be immensely beneficial for some of my research problems. I also wish that they would allow the samples and pedigrees they acquired from Sorenson/GeneTree and used for their new ethnicity estimate to also be included in the current AncestryDNA product. Having those older, legacy samples could really be beneficial.

    • Blaine Bettinger 16 February 2014 / 2:01 pm

      Transfer would definitely be a plus, and I don’t understand why AncestryDNA doesn’t take them, unless they’re truly convinced that their algorithm can’t handle the data. After all, a transfer is almost entirely profit compared to an actual test.

  8. Anthony Parker 16 February 2014 / 10:51 am

    Sorry for the duplicate comment

  9. Bonnie Schrack 16 February 2014 / 11:38 am

    Thank you, Blaine! This must be in the air, you have expressed exactly what has been on my mind. I’m extremely happy with what I’m getting from Ancestry, in particular the matching system with family trees. I understand the frustration for people outside the U.S. I know I’m one of the lucky ones. Many parts of my family came here well before the Revolution and have been well documented. I have 25 leaf hints, and most of them seem to check out well. It’s very interesting to see how certain ancestral couples have a disproportionate number of matches descended from them. This could have several explanations; two of the factors I’ve thought about are that (a) I happen to have inherited more of their DNA than I did from other ancestors, whose DNA didn’t happen to get passed along in the genetic lottery, and (b) they had an especially large number of descendants.
    I’m very pleased with some of the matches, thanks to which I have new connections with friendly cousins who are actively interested in genealogy, something we all hope for! And I’ve been delighted to find that some of the best matches have already uploaded to Gedmatch! This allows me to assign large segments to specific ancestors, way back in my family tree! I share over 50 cM with one woman whose connection with me is through my 4th-great grandfather (we’re working on the question of the gr-grandmother). That may sound large, but I think it’s just a happy random chance that such a large segment has survived. It may be that it could be broken up a little by a different algorithm, but it seems clear that we share a large portion of Chr 11. That’s just one example that I’m a little obsessed with right now.

    My feeling is that while in the past, I held back from recommending AncestryDNA to people who hadn’t tried it, now that I’ve tried it, I will certainly recommend it. I do feel that its usefulness is vastly enhanced by Gedmatch, if you have any serious, ongoing interest in this research, and I hope, as its database keeps growing, that they receive the support they need to keep that site functional and improving.

    • Blaine Bettinger 16 February 2014 / 1:51 pm

      Thank you Bonnie! I’m really excited about the triangulation tool, it could be a great addition to the AncestryDNA product. And congratulations on the success you’ve had, that’s great.

    • Barbara B 18 February 2014 / 10:41 am

      Hi Cousin!

      I think a third factor, in addition to randomly inheriting a big chunk from an ancestor and that ancestor having many descendants, is that ancestor being widely known – i.e., written up in 19th century county histories, etc. A large proportion of the descendants know of that ancestor.

      • Blaine Bettinger 18 February 2014 / 10:51 am

        Barbara – thank you for stopping by and leaving a comment! Knowing an ancestor will certainly be beneficial. However, one power of triangulation and massive data is that the genomes of mystery or unknown ancestors can also be identified and recreated, as long as they had sufficient descendants. That requires more computing power and raw data than ancestors with trees, but it is an example of what genetic genealogy will be able to do.

    • Marci Bowman 24 February 2014 / 1:27 am

      I consider testing at Ancestry one of my truly colossal mistakes. Unfortunately, too many of the so called matches don’t hold up when they do transfer data to GedMatch or when they are evaluated against pesky things like Y-DNA haplogroups. I think that believing Ancestry’s leaf hints and common ancestors can lead one down some very wrong garden paths. .

      Maybe I am just a natural skeptic, but from experience with the profit over quality factor at Ancestry, I think I would bet on acquiring ocean front property in Denver before I bet on Ancestry’s actually checking for significant overlapping segments and triangulating properly.

      • Blaine Bettinger 24 February 2014 / 9:24 am

        Marci – I fear that often much of the frustration with AncestryDNA is due to an underlying misunderstanding of genetic genealogy and the companies’ products. I receive MANY comments and emails about DNA testing, and based on that experience your comment raises a few red flags.

        When you say that the matches “don’t hold up,” do you mean that you don’t actually share DNA at all? That is very unlikely. Any difference you’ve seen is more likely due to differences in threshold settings for AncestryDNA vs. GEDmatch. Determining whether two people share a segment of DNA is a straightforward task. The complications come in phasing the DNA, setting minimum thresholds, and other interpretation of matching segments, not from identifying matching segments.

        Further, there is a chance you are confusing atDNA and Y-DNA. The two are completely unrelated and, in my experience, are only rarely used together. For atDNA and Y-DNA to work together, you have to be examining a related paternal line, that line has to be closely related, and you have to assume that those lines are unbroken. For example, in my own family, other than my father and brothers I would have to go back 7 generations to find another male line with my Y-DNA, and there is only a small chance that we would share atDNA.

  10. Mary E Hall 16 February 2014 / 12:26 pm

    Great post! I use AncestryDNA all the time for some really good clues. In fact, I found a couple instances of “DNA evidence” indicating a possible Virginia connection that I’ve been chasing around for years. I co-lead a DNA special interest group at my local genealogical society and I know that most genealogists struggle over the work involved in interpreting DNA “results”. I’ve been asking myself recently, do we need a chromosome browser? We may not….we just need a way to verify the source information, and what specifically it is telling us, just like any other genealogical based piece of information. Triangulation would certainly help with usefulness.

    Thanks, Blaine, for providing credible notes on why we should all be considering AncestryDNA, too.

    (Bonnie – I, too, have noted that some ancestral couples seem to pass their genes down to more of my AncestryDNA “cousins” than others of the same distance, and considered your explanations, too. But on a couple of them, I know they were in an “endogamous” community of KY/VA (but it could also be Mennonite PA) so am considering that “concentration” as a reason their DNA is so “sticky”).
    Mary E Hall
    Santa Barbara, CA

    • Blaine Bettinger 16 February 2014 / 1:50 pm

      Thank you Mary! I agree that most people would not use a chromosome browser, but that there absolutely has to be some alternative way to verify our matches. I hope triangulation fits the bill.

  11. Kalani 16 February 2014 / 1:20 pm

    Every company has their strengths & weaknesses, just like any other product out on the market. I’ve heard an expert in the genetic field profess how much greater Ancestry is to other companies, but I’ve seen the strength & weaknesses in all. What it comes down to, regardless of all these bells & whistles for any company is what the consumer wants and is looking for. Writing a blog about Ancestry trying to repaint a better picture may help but what should really matter is how each company could be helpful to people, period!

    Only reason why I tested with Ancestry was to see how their “Polynesia” section holds up as I suspected a lot has to do with the fact that Polynesians like other Oceanic people have also elements of Southeast Asian in them. And at the same time, it seems that people who aren’t from that area still come up as showing Polynesia, as was told to me by someone whose father was from the Philippines, but it was a signficant amount, compared to how 23andme shows some Polynesians having very little but I know that is due to some of it falling under the SEA category. But as you see, they all have their faults.

    • Blaine Bettinger 16 February 2014 / 1:47 pm

      While I object to the characterization of this post as “trying to repaint a better picture,” I agree completely that every company has their strengths & weaknesses. But that can be a good thing when it drives the companies to compete, innovate, and improve.

      Regarding “Polynesia,” I would be surprised if any of the companies have a good handle on this, considering the rarity of samples and individuals (which will be true for many regions). At this point I would only recommend relying on academic research rather than an estimate from one of the genetic genealogy companies.

  12. Lloyd Preston 16 February 2014 / 2:28 pm

    The position of Ancestry to exclude non-Americans from testing is extremely frustrating , particularly when they take the position that it is for legal reasons. As a Canadian I have tested with the other companies with no difficulties. Unfortunately nearly all my matches with the other companies are Americans including those with Colonial roots that I am unable to tie together. Being an illegitimate non-American precludes from me from testing and I would agree that Ancestry may be good for adoptees with the caveat that they are American residents. For me I am not going to bother renewing my World Membership as Ancestry is of very limited value considering the subscription cost. The only access I have to Ancestry DNA data is through Gedmatch.

    • Blaine Bettinger 16 February 2014 / 2:41 pm

      Lloyd – I certainly understand your frustration with the lack of international testing. I don’t know why they don’t offer it country by country, as it seems Canada would be an easy one to knock off the list.

      I’ll tell you though, even as an American with old roots, the line “nearly all my matches with the other companies are Americans including those with Colonial roots that I am unable to tie together” applies just as equally to me and most test-takers. The % of matches we are able to verify is shockingly low, but it will only improve with more data and new tools.

  13. Peggy 16 February 2014 / 3:20 pm

    Great post, Blaine, thanks! Also very delighted to hear that Ancestry will be deploying a triangulation tool this year. Are there any other details about it you can share?

    • Blaine Bettinger 16 February 2014 / 9:25 pm

      See my comment below in response to S99H99. There’s not much information out there yet. I have some guesses, but I don’t have any hard information yet.

  14. S99H99 16 February 2014 / 9:02 pm

    Unaware that there was a “witch-hunt” out for AncestryDNA. In looking over the web on genealogy topics I have noticed lately some people pushing FTDNA, but not really trashing AncestryDNA.

    Also, what has been published publicly about AncestryDNA’s promises regarding upcoming tools has been, at best, nebulous, whether coming from AncestryDNA or those who have been to gatherings such as RootsTech.

    “Triangulation” is a term now used for a couple of different processes, and I think we need to come up with better terms to describe specific-segment pools versus set intersections of cousins.

    • Blaine Bettinger 16 February 2014 / 9:22 pm

      S99H99 – Thanks for stopping by! There has been a witch hunt, which you can verify by reading the comments to this post at the ISOGG Facebook group. Many people there echoed my discussion of the “witch hunt” that’s taken place in various places online. It’s hard to deny something that so many people have experienced. Interestingly, however, the AncestryDNA naysayers have been extremely quiet today!

      You’re right, none of us know yet what the triangulation tool will encompass, although I personally heard 3 different AncestryDNA people discuss it at RootsTech. I know from their discussions that it involves triangulation, will probably have a slick interface (perhaps involving “verified” lines?), and will probably not have segment data. Hopefully we’ll hear something more soon

      And I think I understand your last point, but I’m not completely clear on it. I’m not 100% sure that AncestryDNA will specifically use the word “triangulation” to name their tool, but I think everyone knows what I mean when it say it, since it is AncestryDNA’s alternative to segment data. That narrows the meaning considerably.

    • Barbara B 18 February 2014 / 10:36 am

      “Triangulation” that requires a common segment would be great, “triangulation” that just means “In common with” anywhere in the genome, less so.

  15. Claire Tuohy-Morgan 16 February 2014 / 10:49 pm

    I tested with both Ancestry DNA and with Geno 2.0. They actually confirm each other, giving the same basic percentages. What I love about Ancestry is that I am able to match up trees/cousins, and see that some of my ancestry comes from a different genetic population than expected. I am descended from Acadians on one line, and actually most of my Ancestry matches have been through those Acadians. And what they seem to show is that while I show a clear DNA match to these people, many of us show only about 3% Western European genes. I think this must mean our French are descended from other populations. What do you think?

  16. Helen 17 February 2014 / 3:51 am

    I am grateful to those who push AncestryDNA and others to do better, but it would be foolish of me not to fish in all the big pools. I still don’t have any AncestryDNA hints, which probably highlights their lack of European data (all my grandparents were born in Europe).

  17. Michael Ward 17 February 2014 / 4:07 am

    I haven’t seen any sign of a “witch hunt.” Perhaps you’re responding to a small number of people who populate an isolated blog? Ancestry DNA is a wonderful, but irritatingly imperfect, tool for genealogical DNA. I know they are working on an ecstatic solution to all of our problems; but why, in the meantime, do they refuse to give us the simple match data? I don’t need no browser; I just need the numbers. It there is anger at Ancestry, it is because of their “we know best” attitude, and their refusal to communicate with their users.
    Software is hard! It’s not as easy to do a general match-by-segment as they thought it was going to be. OK, OK, give us what you do have, and we’ll help you look forward to your promised nirvana. Meanwhile, just the chromosomes and the start and stop points … and Stop Pretending Nobody Is Asking for This.
    Of all the things that Ancestry.com does, their Stonewalling is the thing that irritates me the most. Embrace your customers, and make them your helping agents!: basic law of commercial success. Not in Ancestry’s genes, sadly.

  18. Lynn David 17 February 2014 / 5:54 am

    I’ve tested with all three and find the 23andMe and AncestryDNA ethnicity tests are somewhat in line. Although 23andMe had me listed as less ethnically Great Britain than AncestryDNA but after AncestryDNA Ethnicity v.2 came out for some reason 23andMe also bumped up my British ancestry (and they identified that 0.3% Asian of mine as “Yakut” at the same time)..

    But I do have a beef with Ancestry’s ethnicity measurements. But I’m trying to help and I told them about it here: http://ancestryforums.custhelp.com/posts/cd43442e65

    I think I have certainly pointed out a problem they have in evaluating the ethnicities in Europe. Especially for those of us with NW German Saxon or Belgian Wallon bloodlines (where BTW, I discovered an 8th cousin via 23andMe and our two trees – sometimes those 0.08% matches do mean something). My analysis may be a simplistic, but I was still a bit peaved about one reason they gave concerning keeping their British ethnicity so broad. It was: “individuals that make up our Great Britain region do indeed have deep roots in Great Britain, and so is consistent with the naming and definition of all of our other regions.” Yes, their roots go back to their Saxon immigrant ancestors into Britain. But mine likely do not (I’ve not heard of any migrations of British Saxons back to their ancestral lands).

    When AncestryDNA Ethnicity v.1 came out the explained the overabundance of Scandanavian ancestry by Ancestry was: ‘that represents your deep roots.’ And now it seems that Ancestry has gone the other way and chosen to identify the shallower roots of those British formerly classed as Scandanavians and have left out the middlemen – the German Saxons.

    This reminds me of my beginning in genealogy back in the late 70s when I first decided to get into it. I found that in reviewing the local libraries that there was a prevalence of Colonial American and British genealogical sources. I understand that bias, and it made me a better genealogist for having to delve into the first-hand records – to take my family history into my own hands. But the attitude towards pleasing those of British descent seems to me to be that old genealogical bias among Americans rearing its head. Which is more correct, to place most of those with known ‘shallow’ British ethnic ancestry and at the same time suck in Saxons and other continental European ethnicities into a British ethnicity for which they have no affinities. Or identify these people as western European or undifferentiated western (or northern) European and have those same people with British roots realize that it is then their deeper (Saxon) ancestry which is being determined. Frankly, I think the latter.

  19. GC 17 February 2014 / 7:48 am

    I think it is just the general impression that Ancestry is a rip-off outfit—i.e. mercenary…
    Why Ancestry DNA?
    FTDNA is surely the biggest & technically the best…at any rate?
    Rgds
    GC

  20. Blaine Bettinger 17 February 2014 / 8:06 am

    GC – congrats on having the least productive and uninformative comment on this post so far! And that’s out of 36 total comments! You are exactly the type of person I wrote this post for, and I have a nagging suspicion you didn’t even read it. And Roberta’s post is almost 2 years old, which in the world of autosomal DNA might as well be 20 years old.

    All the companies have pluses and minuses. FTDNA is an INCREDIBLE company, and I’ve tested there many, many times. I’ve recommended that people test there many, many times. But telling people to avoid AncestryDNA is foolish.

  21. Anastasia 17 February 2014 / 9:59 am

    Blaine, I have had the same perception. All the companies have vast room for improvement, and all offer a valuable service. It does seem that Ancestry’s product’s cons are more heavily publicly ostracized. At any rate, thank you for this post.

    Anastasia

  22. larry thompson 23 February 2014 / 4:06 pm

    I read a report about 20 years ago that estimated that in about 20% of cases the husband is not the father of a child. If true it’s going to drive a lot of people nuts over the result of their dna testing.

  23. Elizabeth 12 March 2014 / 2:51 pm

    While I agree that AncestryDNA gets a lot of negative attention, I don’t think it’s a “witch hunt” at all. I think there are reasonable explanations for why it does get so much criticism.

    The first is that it’s the largest and most well-known genealogy company, so it’s bound to attract the most attention, whether positive or negative.

    The other issue is that AncestryDNA is marketed to people who rely only on paper trails and family tree comparisons. As result, its average user, as you’ve pointed out, really has little-to-no understanding of genetic genealogy; they don’t necessarily understand that what shows up in your paper trail and what shows up in your DNA are often two completely different things.

    So, criticisms of Ancestry’s DNA test in the latter area *are* often unfounded. I just happen to think it’s their own doing. With no focus on analysis and the very superficial way in which they market their tests, how can they complain? Ancestry treats its DNA tests like offshoots of its main focus (trees and paper resources), so they’ve attracted customers who use them that way. AncestryDNA can’t expect to attract customers and criticisms from people who understand genetic genealogy because they themselves offer no way for customers to understand their genetic data or how its relevant to their trees. No one at Ancestry can now be shocked and indignant that they’re getting comments like: “This is a scam! My tree goes back hundreds of years in England and I’m not showing any British DNA!” or “I bet they just looked at my tree to come up with these results and matches!”

    Then, they also get criticisms from people who *are* into genetic genealogy. I think they’re completely warranted. If you’re going to show us matches connected to family trees, let us see *how* we match those people so that we can move forward in our research. Many of those trees (most of them, for me) are useless without additional tools. There’s no excuse for that.

    As an African American (admittedly, a small percentage of their customer base – but a customer, nonetheless), my need for that segment data is more urgent that average. Firstly, lots of my black cousins will have no surnames in common with me because our enslaved ancestors’ families would have been separated frequently. After emancipation, our ancestors took on a variety of surnames that may be different even from those of their own parents. They may not have known where their family members were or even *who* they were. This means that eleven siblings with the same biological parents can all have different surnames. We need more than just trees.

    Secondly, since whites test at a much higher rate than blacks, the vast majority of my matches are white. A pretty significant percentage of my ancestry is actually European (which is true of most African Americans, although most of us don’t know that until we test). Turns out that just over a quarter of my DNA is European – all coming from the era of slavery. I don’t have any white grandparents or great grand parents. A few of my 2nd great grandparents – all former slaves – may have had white fathers. Slave children of white fathers rarely had the white father’s surname. Without segment data, almost all of my matches’ trees are useless to me. I can find lines that I *likely* descend from, but there’s no way for me to know: A) which repeat lines among my matches are purely coincidental or B) which ancestors/lines of descent I should be focusing on.

    Offering tools that allow us to analyze our data will go a long way in changing the way Ancestry is perceived and the type of feedback it receives on its DNA tests.

  24. Sharon Remmey 20 June 2014 / 6:23 pm

    I want to do autosonomal testing on husband and my families. I am interested in our Mom and Dad’s families. As I read all these comments I am totally lost. Once the testing is done are there groups you can join to get help. Maybe local public libraries or other groups. I am afraid to pay for testing and then have no clue what I am supposed to do from there. I can prove my Mom,s paternal side to the American Revolution. Any suggestions that could help me would be appreciated.

    • Doug Biggs 5 March 2017 / 12:29 pm

      Sharon, I build a tree on Ancestry.com and took a DNA test. If you haven’t started, then I suggest you start there. I found their finger tip access to Census reports, marriage licenses and such to be helpful and consumer friendly.

  25. Danette 21 June 2014 / 10:56 am

    Yeah it would be helpful to know what a segment data is and how it would help me …its become very technical more so than I had any idea it would be.

  26. Arthur 2 July 2014 / 11:10 am

    I am trying to choose which company to do my first genetic genealogy test, and my budget is limited. I can trace some family lines back over 500 years, but in my paternal (male-only) line I know nothing before about 1800. Family is European Jewish. Should I start with FTDNA? or one of the others, what do you suggest

  27. David Smith 15 July 2014 / 8:13 pm

    Unfortunately some of the biggest AncestryDNA bashers are “23&Me Ambassadors” who, of course, never get around to mentioning that in their posts.

  28. Chris 26 July 2014 / 9:09 am

    What I did was test myself, wife, mother and father when there wa a sale at ancestry.com. I then downloaded the results and uploaded to FamilyTreeDNA (for a fee) and gedmatch.com (for free) This gives me three large databases to compare. Neither 23andMe nor ancestryDNA allow uploads. It looks like ancestry and FTDNA use the same chip (Illumina OmniExpress) for testing so the results should compare well.

    I did find some matches that are from the same towns in Ireland that my ancestors came from. One town Frenchpark, Co Roscommon has a population of about 200 so there is a small gene pool there. That person matched both me and my mother.

    Both FTDNA and gedmatch have tools to look at matches. The Tools at FTDNA are easier to use.

    I admit that most of my family tree was built by searching records but I did connect with a few people using DNA testing that I might not have. We shared an interest in researching. So in the end I owuld say the testing was a good way to find connections to research further.

  29. Felix 30 July 2014 / 1:47 pm

    I have to add to the list the inability to search your DNA matches as a huge disadvantage as well. If I come across a person’s tree where we suspect a relationship, they give me their test name. I should be able to go directly to my results and search for their test name, but I can’t! Surname searches don’t help in this matter bc I suspect that my gggm is the sister of their gggf and neither of us know who the parent’s were. Seriously, I can’t even search my list without a chrome plug-in!

    Sorry, but Ancestry.com DNA has been useless for me. Just a list of names with predicted relationship and a link to their tree only works if you and your matches have a complete tree that is accurate. Very few of us have that!

    I guess I’m one of those because I have driven several people away from Ancestry.com….not by bad-mouthing them, but laying out the facts. This is what you get with Ancestry, this is what you get with FTDNA, and what you get with 23andme….you make the choice.

  30. Dorothy Coltrin 2 September 2014 / 5:15 pm

    I am pleased to see people starting to use DNA to confirm their genealogical paper research. However I have gotten messages from people who now don’t trust Y-chromosome and think they can use their autosomal DNA matches from Ancestry.com to prove their connection.
    I am working with a very common name…more than 150,000 have this name today. My belief is that these individuals WANT to belong to a certain group with this name. So they distrust Y-chromosome and have found some sort of match with Ancestry’s autosomal!!!
    I think Ancestry needs to let people know that there many many false positives using autosomal DNA, especially with a common name. And I become frustrated with the Ancestry trees that are connectied…with no documentation or just copied from someone else who has limited paper evidence.
    Autosomal DNA is a wonderful tool for genealogy, but it has its drawbacks.

  31. Marsha Rice 6 October 2014 / 3:16 pm

    Gedmatch is a great tool for those who have done ancestrys autosomal test but want more tools to work with. You just have to zip your raw data from ancestry and upload to gedmatch.

  32. Lynn 19 October 2014 / 5:42 pm

    Too bad Ancestry dna testing is not available to Canadians. Wish it was.

  33. Jon 7 April 2015 / 4:32 pm

    I’ve been involved in traditional genealogy for several years, deeply involved for just over a year. I started doing DNA testing via all three of the major companies about a year ago as well. I’ve found Ancestry’s information to be a helpful tool – and like all of ancestry, I understood that I needed to apply a little logic and healthy skepticism to anything – rather than blindly trusting any results. With that approach, I was able to find some great insights and was even able to get through one particularly nagging brick wall I’d been facing.

    I actually avoided the more technical side of genetic genealogy because of the very thing you’re referring to. I’m actually a very technical person and enjoy getting into the “guts” of tech and research. But I was extremely turned off every time I wanted to learn more about how to use chromosome browsers or to understand how to strategically approach segment data. It seemed so many people were bashing Ancestry – which had proved to be a useful too. Then they were even fighting amongst themselves about FTDNA and 23andMe. In retrospect, I understand their viewpoints, and with a few exceptions, see why they had such strong opinions. But within our community, I think we’re doing a disservice to our future if we simply bash Ancestry when it is proving to be the most effective way to get people to join us in our quest.

    When someone asks me for a recommendation, if we’re talking autosomal testing, I generally recommend they test with Ancestry so they’ll have exposure to a huge database and some pretty easy-to-use tools AND that they then transfer their data to FTDNA for more technical results.

    Personally, If we’re going to bash people, I’d rather bash people on FTDNA and 23andme that don’t respond to requests, or who post no supporting information. We’re all in this together and a rising tide lifts all ships.

  34. Laura 9 September 2015 / 10:48 pm

    My results came back within 2 weeks not 6-8. I question the accuracy the race mix is and how my entire family looks. Did they mix up my test? Wasted money and I am left disappointed.

  35. Karl 8 December 2015 / 4:55 pm

    Ancestry DNA has SERIOUS issues. The search results can very day-to-day. For the first week of my membership Ancestry could not get the Search function to work properly. The first couple of days- ZERO results. After a few days- slowly getting results. After 8 days- finally what I thought were proper results. But it’s wrong…. While I spent a week waiting for Ancestry to figure out just how in the world my account was messed-up (including wiping every memory… wiping my browser history… wiping every cookies… clearing the cache…. oh, and trying it on 6 different PCs with 5 different operating systems and several browsers/versions of browsers) I went nearly page by page through my 120 pages of results. Sure I jumped around in chunks, but I still reviewied MANY pages and trees manually. I noted some of the occurences of my surname and others of importance. I even did a couple of screenshots for the more interesting ones. However, now that Search is supposedly working correctly and fully, I can’t find some of those interesting results. They are NOT in my Search results. And since I was only interested in snippets, I didn’t bother to note their name while in manual review… So, do I start going through all (now) 122 pages one-by-one looking for relatives because Ancestry cannot be trusted? How many people have NO clue their results are not accurate or complete?

  36. Cal Crowley 17 July 2016 / 1:07 pm

    AncestryDNA needs to supply the chromosome data to prove to me that they are not JUST looking at family trees to decide Hints matches. I had two just recently that I suspected they added to my Hints just because of our tree data — AND NOT BECAUSE OF DNA data. Unless I can see the chromosome data myself to look for shared overlaps I have no warm fuzzy that AncestryDNA is correct about the match. Its really that simple. And speaking of trust: since Ancestry dropped support for FTM I do NOT trust them at all to be there for me, honestly.

  37. Barbara 11 January 2017 / 10:06 am

    I have tested my family and thinking to hire a genetic genealogist, but I am concerned with the privacy issues. How does it work? Do I need to provide the log in information to the genealogists and he/she will gain access to my account and my raw DNA results or there is another way to handle it?

  38. Chris 23 April 2017 / 8:10 am

    To get the greatest coverage database wise we tested on Ancstry, downloaded raw data, and uploaded to FAmilyTreeDNA and gedmatch. Now we have three databases of data and ccan use the tools on gedmatch and FTDNA.

  39. SThomp 22 December 2017 / 12:20 am

    Whenever I’ve put a quickie tree on Ancestry I make it private and share it directly to only those who I’ve told about its limitations.

    I wish Ancestry would also allow us to make only parts of trees public. Most of my trees are private because I don’t want people copying info that I’m not yet quite comfortable with the degree of likelihood that it’s accurate. Yet, I’d like to widely share the parts of my trees showing those ancestors I’m sure about.

Comments are closed.