Tuesday, December 18, 2012

Comparing Admixture Test Results Across Companies (otherwise known as "ethnic" breakdowns): FTDNA, AncestryDNA, 23andMe and Geno 2.0 - My Review

With the recent releases of Geno 2.0 and the new 23andMe Ancestry Composition feature, I thought it would be a good time to review and compare my biogeographical ancestry (BGA) results from each of the major DNA testing companies. (BGA is sometimes referred to as ethnic breakdown or admixture analysis.)

My Known Ancestry
First, let's take a look at what I know of my ancestry:
25% Finnish
12.5% Scandinavian (Norway)
12.5% British (England)
~25% Colonial New England (presumably mostly British)
~15.63% Unknown (due to brick walls and immigrant ancestors of unknown origins)
~9.38% German

CeCe's Ancestry Percentages
You can create your admixture pie chart here.

Family Tree DNA's Population Finder
FTDNA's Population Finder (comes with Family Finder) was one of the first commercially available admixture tests based on chip technology.

I am listed as 98.93% Western European (Orcadian) with a 1.31% margin of error. Surprisingly, none of my substantial Finnish ancestry was detected. In my case, Population Finder is not very informative so I am glad that FTDNA will soon be updating this feature. Of course, many people's Population Finder results will be much more detailed than mine, so my experience may not be a typical representation of this feature's capabilities.

AncestryDNA's Genetic Ethnicity
Just a couple of weeks ago, AncestryDNA's admixture tool was the newest entry into the market, but times are changing fast.

It's one thing for results to not be especially informative like Population Finder above, but for them to be downright wrong is disappointing. This feature underestimates my Finnish ancestry and overestimates my Scandinavian ancestry, both substantially.  Because I have tested my mother, I know that they are estimating a large amount of Scandinavian from my mother's side when there is no Scandinavian ancestry in her extensively documented family tree at all. Does that mean that she doesn't have any? Not necessarily, but AncestryDNA estimates that she has 62% Scandinavian and only 23% Finnish (instead of ~50%) ancestry. Although none of these admixture tools are perfect, this one seems to be the most misleading. Notably, I have received many messages from people confounded and even very upset by their AncestryDNA Genetic Ethnicity results.

23andMe's Ancestry Composition
23andMe started out with a very conservative three population breakdown called Ancestry Painting. Customers had long been requesting an updated version. Happily, the brand new Ancestry Composition is much improved. There are three different views and a number of resolution and confidence settings. I prefer the most granular (Sub-regional Resolution) and lowest confidence (Speculative) of those.

Map View:

Chromosome View:

Split View:

This version matches my known ancestry the most closely and is also in line with the ancestral origins represented in my Relative Finder/Family Finder/AncestryDNA match lists. It was fun to see that even my tiny amount of Ashkenazi Jewish ancestry was detected. Although this tool is definitely the most informative and flexible, it isn't without problems in this early version. For example, on the Split View, all of my Finnish should be from my maternal side, but you can see that 4.6% of my estimated 21.5% Finnish is erroneously assigned to my paternal side. Conversely, almost all of my Scandinavian is correctly assigned to my paternal side. Since my deceased father wasn't tested at 23andMe, they had to compute the Split View without my father's results, which makes it more challenging. However, since I have tested both of his siblings there (they don't have any Finnish), advancements in the future may help in this regard.  I have many relatives tested at 23andMe and the vast majority of them have results very much in line with their family trees. Both of my sisters' Ancestry Composition results seem even more accurate than mine, as do my mother's, both of my aunts' and my uncle's. (I may just be a bit of a mutant!)

National Geographic Geno 2.0's Who Am I?
As most of you know, the new phase of the Genographic Project started returning results a couple of weeks ago.  Screen shots show the different components that make up my ancestry according to these results.

Then, they list the two reference populations that match the combination of my components the most closely. Interestingly, even though I am only one quarter Finnish, they chose Finnish as the first one.

Next, they choose Iberian (Spain and Portugal) as the second population, which doesn't fit with any of my known ancestry.

This admixture tool is different than the other ones reviewed here because it is looking at deep ancestry from thousands of years ago. Personally, as a genealogist, I don't find it as informative or as satisfying as the admixture results that reflect more recent ancestry. I also don't really like how each person's unique combination of ancient components is essentially forced into a single (more recent) population label. This, no doubt, works very well for those whose ancestors have been in one geographic region for hundreds or even thousands of years, but it does not result in highly relevant results for those of us who are highly admixed with ancestry from many different regions. I find the tools that analyze and attempt to categorize small portions of my DNA separately, giving percentages of each region or ethnic group that add up to my total ancestry more meaningful.

At this point in time, for those interested in the origins of their ancestors in the last couple hundred to few hundred years, 23andMe is the best choice in my opinion (especially at the new price of only $99). If your interests are more focused on deep ancestry and anthropology, then Geno 2.0 would be my recommendation. (Remember though that this review is focused on admixture features only. There is much more to take into consideration when purchasing a DNA test for genealogical purposes!)

All of these companies would be well-advised to focus on improving their admixture results for those with non-European ancestry. (Obviously, the majority of their customers have been of primarily European ancestry so far, but optimally everyone will have the the same opportunities of self-discovery in the future.) 23andMe has publicly stated their plans to do so and I'm sure that the other companies will address these shortcomings as well.  AncestryDNA's acquisition of the Sorenson database and National Geographic's vast collection of indigenous population samples puts them in the best position to advance in these areas in the near future, but as 23andMe has demonstrated, customer samples carefully vetted with genealogical information can be utilized with great success as well. As more people are tested, particularly internationally, there will be improvement across the board.

It is important to remember that as intriguing as these admixture predictions are, none of them are 100% accurate at the granular level. We still have a long way to go before anyone can honestly claim to be able to tell a person exactly where their ancestors once lived based on their autosomal DNA alone. However, we are making progress and in a couple of years we will likely be amazed at the advances. In the meantime, the ever-increasing competition between the major companies is proving to be beneficial, spurring all to improve their offerings. So, don't count any of them out quite yet. In a short time, my opinions regarding the "best of" may have completely changed again!

*For more information on these features, please see:

[Disclosure: I have relationships with all of the companies reviewed above. I received complimentary Geno 2.0 and AncestryDNA kits for review purposes. I occasionally receive complimentary 23andMe test kits through my work as a volunteer "23andMe Ancestry Ambassador". Additionally, I have affiliate relationships with Family Tree DNA, 23andMe and AncestryDNA and own a small amount of Ancestry.com stock. I always strive for complete objectivity - none of these relationships have affected the opinions stated in this blog in any way.]


  1. Just curious, have you run your autosomal through Doug McDonald? How did that fair?

    1. Yes, in 2010. Good question! I only reviewed the commercial offerings, but third party admixture tools are certainly relevant as well.

      Doug wrote:
      "Europe 97.1%
      Oceania 1.0%
      America 1.5%
      Africa 0.0%
      E. Asia -0.0%

      Europe 91.9%
      S. Asia 8.2%
      Oceania 0.7%
      America 1.1%
      Africa -0.3%
      E. Asia -1.9%

      Europe 89.0%
      Mideast 4.1%
      S. Asia 7.7%
      Oceania 0.7%
      America 1.2%
      Africa -1.0%
      E. Asia -1.9%

      This is classic all-European, specifically British."

      When I asked if he detected my Finnish, he answered:

      "Actually, no. I would see it if the Finnish were mostly Saami, but
      even at 25% there are plenty of Finns that are close enough to
      British to not show up. The same applies to very very northern
      Norwegians, who are also part Saami, but southern Finns are themselves
      very hard to tell from British."

      Thanks, Attica!

  2. Thanks, CeCe, for a helpful and informative review.

  3. Thanks for the review! I hope to see changes at FTDNA as that is where I have tested most of my relatives.

  4. Great post CeCe. One thing is for sure, we've had a lot of rapid change and I hope we can look forward to more refinements in the next years so that you (and the rest of us) get the opportunity to do this all over again:)

  5. Another thorough and fair set of reviews. Could the fact that Finland was governed by Sweden (and also by Russia) for long periods of time in past centuries have stirred up and re-distributed Finish DNA? The 17th or 16th or ? century ancestress that you and I apparently share apparently had a female descendant in New Sweden around 1700. Other matches still living in Finland add evidence that she was ethnically Finish but also documents a little stirring of the genome pot.

  6. Interesting :-) I have not been tested at Ancestry or geno2.0, but I agree 23andme's is closer.

    My known ancestry is (approximately),
    64% German (includes Alsace-French)
    29% Irish
    6% Scottish
    1% English

    23andme has me as (speculative),
    32.5% British & Irish
    17.5% French & German
    0.6% Scandinavian
    0.2% Finnish
    41.6% Non-specific Northern European
    1% Iberian
    1% Italian
    2.5% Non-Specific Southern European
    3% Non-Specific European
    0.1% Middle Eastern
    0.1% Native American

  7. CeCe
    Great feedback. I totally agree with your assessment. I suspect that FTDNA and ANCESTRY will improve as they roll out their newer versions. But for now I would recommend 23andme's over all the rest for Ancestry composition. I have seen quite a few of the new Geno2.0 and I am somewhat disappointed overall. The potential is there to greatly improve over time but the initial results are in my opinion worse than ANCESTRY's. I also have concerns about the reliability of the YDNAs SNPs with some of the problems being reported.

    However all the companies have their periodic growing pains. Overall good things have come the past year. hopefully 2013 will be a good year as well!
    Happy Holidays

  8. Concerning your Geno 2.0 Iberian match, could it be that among your unknown immingrant ancestors there were some Spaniards/Portuguese or people coming from those territories in the New World? Although it's also possible that Geno 2.0 simply chose that population for whatever reasons. I've read that they have made some changes later, as in Auntsha's case.


    1. Hi Luis,
      It is certainly possible, however I don't have an RF/FF/AncestryDNA matches to people from this area of the world, so it would have to be pretty far back and likely not a substantial component of my ancestry.
      I really don't think that is how the test works though. It appears to me that they are trying to fit a person's overall ancestry into one reference population and mine simply cannot be categorized that way.
      Thanks for your comment!

  9. CeCe,

    You may be right, that is what they seem to be doing: tryig to fit a persons overall ancestry into one of their referenc populations. Althoug that may not work sometimes and they will have to look for a better fit, I guess.


  10. Great review and comparisons of all the the companies. As far as the new Ancestry Composition I find the Speculative the closest percentages to what I have figured my ancestry is. From all other tools my unknown half is probably Southern Europen and area around Italy. They were the first to pick up the Italian at 28% the Iberian of 3.3% from my maternal Canary Island and Valancia side and 16% or so non specific. My Northern has a lot of non specific at 25% or so and they did pick up Irish/Brit but still very low% and French/German at 9% (should be 18% Fr) On the standard setting the % 's are just too low and too non specific that it makes it look like they really don't have a handle on analyzing DNA! I am just happy that all this competition is making each company want to put out a better and more accurate product! It definitely is a win, win situation for us genealogy junkies!


  11. “What exactly does British Irish ancestry mean in the Ancestry Composition?”

    I am curious what actual group constitutes "British" ancestry. It was my understanding that England is actually a melting pot of many groups. Does the new ancestry composition tool refer to the more ancient DNA or invading groups, such as Angles, Saxons, Vikings, Normans? I realize that the tool is supposed to reflect the last 500 years. What does it really mean to be British? How are British and Irish the same?
    To illustrate, my known genealogical ancestry :
    1/2 German (late 19th century, half from Oldenburg in Lower Saxony)
    1/4 Colonial American (mostly English, but Mom's maiden name is Irish--haven't proved)
    1/8 French Canadian (1640 arrival in Canada, paternal grandmother born in Canada)
    1/8 English Canadian (fairly recent-- Cottingham )

    Given the above, imagine my surprise when Dr. McDonald's report simply stated:
    "100% English and little else." !!! I"m 1/2 German!

    Well, via further communication, it was explained to me that my relatives from Germany were from Lower Saxony, whose people were Saxons who invaded English. Most of my French relatives were from New Rochelle-- another invading group.

    My speculative Ancestry Composition results:

    100% Northern European

    35.3% British and Irish
    14.6 French and German
    2.6 Scandinavian
    44.9 Non specific Northern European
    1.0 non specific eastern european
    1.6 non specific European
    0.0 Ashkenazi

    My mtdna group is T1a1

    When I was in Italy this summer, I was repeatedly asked if I was Scandinavian as I am tallish, fair, with reddish dishy hair and round green eyes, long bony nose.

    Sorry lots of info and questions, any insights are appreciated

  12. I disagree Cece----Price is not everything & your criteria for reaching this conc. is let's say skewed

    1. You are certainly free to express your opinion. Geronimo. However, I never said that price was the deciding factor, but it sure does make it more accessible for many people who might otherwise not be able to participate.
      So, do tell, which commercial admixture feature do you like the best?

  13. Thanks for the review! Based on your recommendations, I picked up two 23andMe kits as Christmas gifts for my parents who want to learn more about their ancestry.

    1. A Belated Merry Christmas to them (and to you)! I hope they enjoy your thoughtful gift.

  14. I also thought that was a great article that I mostly could follow, so based on your recommends I ordered 3 kits from 23andMe, two as Christmas presents, one for a granddaughter who doesn't even know who her father is (as I didn't long ago), and one for my son who is yearning to discover more about both sides of his family. The third test is for me, to more correctly verify what my Ancestry tests have told me so far. I also appreciate the lowered costs.

    1. I hope that you all enjoy learning more about yourselves through DNA testing. With the increase in the size of the databases, a lot of people are solving mysteries in their close families. I hope your granddaughter will be one of them.
      Thanks for your comment,

  15. This may sound funny but my grand daughter and her father have inCREDibly long arms and I can't help but wonder re: Neanderthal, which I have seen 23andme discuss. Does this test include this subject?

    1. ;-)
      23andMe and Geno 2.0 both give percentages for Neanderthal. However, they seem to differ and I'm not sure which is more accurate.

  16. I appreciate your honesty and price certainly does matter to some of us. Have you done a similar comparison for mt dna tests? I have been a bit disappointed with ftdna for autosomal so would like to try a different company and wonder what your opionion is about mtdna testing. Thank you.

    1. Hi,
      Thank you.
      I have not done a comparison of mtDNA tests because, in my opinion, there is really only one place for that - FTDNA. (This is also true for Y-DNA tests.) They are far and away the leader for these tests and the only place that offers full mitochondrial sequences to the public. Ancestry.com's offerings, results, tools and support are dismal in comparison. (Yes, I ordered a mtDNA test there before I knew any better, so I speak from experience.) There are other smaller companies who offer various Y-DNA and mtDNA tests, but I wouldn't recommend any of them in comparison to FTDNA.
      As far as mtDNA for genealogy, I usually don't recommend it because it isn't usually very helpful. If you are going to do it, the full sequence is most likely to be relevant to your genealogy. My case is unique because my mtDNA was informative. Further info here: http://www.yourgeneticgenealogist.com/2010/12/my-full-mitochondrial-sequence-and.html
      We were very surprised to learn that a mtDNA test did solve an adoption search recently, so you never know.
      If I were you, I would give autosomal DNA another shot by testing at 23andMe. There is unlimited potential for discovery with atDNA, but it depends on a lot of work and some luck.
      Thanks for your comment,

  17. And thank you for helping average people figure out this exciting albeit complicated and somewhat confusing process.

  18. CeCe, being relatively new genealogical research I had heard of but not really looked into DNA testing. Your article has really piqued my interest and I will now pursue. Thanks for your contribution to all our research endeavors.


  19. I'm studying your review of autosomal DNA tests with interest. I ordered AncestryDNA tests for myself and my wife, and both sets of results presented surprises. However, I'm curious about your critique of your results from AncestryDNA--that they got the Finnish and Scandinavian portions wrong. Please tell me if my following remarks are valid.

    1. It's hypothetically possible (however unlikely) for two siblings to receive completely different autosomal DNA from their parents: that is, sibling #1 receives half from mother and half from father, but sibling #2 receives the other half from each parent.

    2. It's therefore possible for siblings to receive significantly different results in genealogical DNA tests.

    3. As far as ethnicity goes, it's possible for recombination to, in effect, erase portions of your ancestry and, in test results like those you discuss, present others in percentages that are high in relation to the documented family history.

    For example, both of my grandmothers had well-documented completely British backgrounds. Approximately 18% of my ancestors came from Germany and Switzerland, part of the area AncestryDNA calls "central Europe." One of my grandfathers was an orphan whose origins are obscure, so there's a 25% "wild card." The remaining 7% is probably British. However, my AncestryDNA test results show 71% central Europe and 29% British. If I understand chromosomal recombination correctly, the British portion diminished by random assignment over a couple of generations, while the central European portion happened to survive recombination to a greater extent.

    However, this is a new field of study for me. Could you comment?

    Thank you.

    1. There is an average of DNA that we are expected to inherit from each level of ancestor. In reality, it varies and the further back you go, the more likely it is to be lost. However, at fairly close relationship levels like my ~100% Finnish grandmother and I, the percentage of DNA should be pretty stable. We all will share approximately 25% of our DNA with each of our grandparents. It could be a little less or a little more, but not wildly so. Therefore, it is unreasonable to think that I would have received much less than 25% of her Finnish DNA and, instead, received a larger share of Scandinavian DNA from my Norwegian great grandmother (should be ~12.5%). From my extensive experience working with this data, I know that not to be the case. 23andMe was able to distinguish my Finnish DNA quite well and my mother clusters nicely with the Finns in other PCA plots I have seen. My match lists are FULL of Finns, so much so that I would have theorized that, if any variance existed from the average, I inherited more than the average amount of DNA from my grandmother.

      Here is the average DNA sharing/inheritance that we expect to see from our ancestors:
      Parent = 50%
      Grandparent = 25%
      Great grandparent= 12.5%
      2nd Great grandparent = 6.25%
      3rd Great grandparent = 3.125%
      4th Great grandparent = 1.563%
      5th Great grandparent = .781%
      6th Great grandparent= .195%
      7th Great grandparent = .049%

      As you can see, the amount that you receive from your ancestors quickly diminishes as you go further back in your tree. Therefore, for a more distant ancestor, it is certainly possible that you would not inherit any DNA from them and thus their ancestry might not show up in your DNA analysis at all. However, unlike cousin matching, these admixture analyses use very small blocks of DNA not the larger ones that cause a predicted relationship to appear on your match list. So, it is likely that we are seeing contributions from ancestors very far back in our trees.
      I am going to post this and then answer your points.
      (to be continued...)

    2. 1. It is theoretically possible, but never happens in reality. The lowest sharing I have ever seen between two full siblings is about 39%. It may go a bit lower in practice, but not much.
      2. The odds of two full siblings not only receiving substantially less than the average DNA in common, but also of each sibling receiving specifically those portions of each of their parents' DNA that have diverse origins has got to be astronomical. I'm not saying that siblings don't see differences in their ancestral breakdowns - my sisters and I certainly do, but it will rarely be authentically significantly different.
      3. Yes, this is true, but with the limitations already discussed. For very small portions of ancestry this can certainly be the case and I see it frequently.

      As to your example, your explanation could be plausible. However, if customers of AncestryDNA weren't constantly having to make these leaps to explain their Genetic Ethnicity results, I would feel more comfortable accepting it. I will be very curious to see what the third party tools make of your ancestry once AncestryDNA releases the raw data. Please check back with me at that point if you wish to update me (and my readers).

      Thanks for your comments.

  20. To follow up on my earlier post with the three numbered remarks, here's another way to express what I think I'm understanding: a DNA test shows not your ancestry but your array of chromosomes. Some chromosomes survive recombination, others don't. The DNA test shows your unique chromosomal identity, which represents a random selection from your actual ancestry but doesn't accurately reflect the whole.

    Does this make sense?

    Thank you.

    1. Yes and no. Your chromosomal identity and your ancestry cannot be mutually exclusive. Although we are each made up of a unique genetic combination of our ancestors, the inheritance pattern still follows certain rules. It is not so random that it defies logic.
      Thank you for the discussion.

  21. I received my Geno results Jan. 3rd, however, I'm not very impressed. My hg was changed from R-P312 to R-P310 or R1B1A1C. I've read this is because the current Geno chip does not detect R-P312. Is this true? Also, my first population was Danish and the second one, Tunisian. I guess I should interpret it as saying my ancestry is somewhere between those two countries? I don't think many people have Danish as their first or second population. But, anyway, I just look at it and say 'interesting'. Also, my Neanderthal score was 2.1% vs 23andMe's 2.7%. My Denisovan (God knows what that is) is 3.9%. I don't take this score seriously, as they are still experimenting with it.

  22. I have tested with FTDNA, Ancestry.com, and 23andMe. The first two were very similar in their findings of (+- a few points) 47% European (Eastern and Southern), 33% Native American (North and South), and 16% North African/Middle Eastern. 23andMe was relatively close with the first two but showed less than 1% on the North African/Middle Eastern. That much of a difference seems to be huge! It feels like a whole different person...

  23. Hi CeCe and everybody, I took Geno 2.0 and my score was largely as expected, I guess, being a mixed Swedish speaking Finn/Finnish speaking Finn from Finland. The "logical" shares were 50 % Northern European, 23 % Mediterranean, 19 % Southwestern Asian and 5 % Northeastern Asian. But on top of that I also scored a very nice but equally mysterious 2 % Oceanic and 2 % Native American!!! My reference populations were British (UK) first, and Russian second. And finally Neanderthal 2 something % and Denisovan 3 something %. How should the odd Oceanic and Native American elements be interpreted? I have now knowledge of that kind of genes in my ancestors. Could the analyses simply be flawed? Ideas, CeCe, anyone?

    1. I think your "Oceanian" admixture could be something old and unknown, perhaps from Siberia, that the tools identify as Oceanian. I am a Swede with, as far as I know, only Swedish and Forest-Finnish ancestors and I sometimes get smaller amounts of mysterious "Oceanian", when using the admixture tools on GED-match.

    2. Now I have discussed with others. I seems most likely that your "Oceanian" as well as your "Amerindian" admixture is an ancient Eurasian component.

    3. Look here:

    4. part 1:
      I was wondering if anyone has any updated thoughts on this?

      My mom, of Latvian heritage (some parts are probably extremely long and deeply rooted in Latvia, other parts probably eventually go to Sweden/Germany/Netherlands/?), had a shocking 2% Oceanian component show up too.

      Some basics:
      my mom's Geno 2.0 gave 2.1% Neanderthal and 3.0% Denisovan
      (current avgs they list as 2.1% and 2.1%)

      (Latvian heritage; thought to perhaps be considerably deep on grandfather's lines back (although the family trees is barely known at all on this side, not past his parents, but there was always talk for some reason of deep and long pure Latvian roots on his side); going back to nearly the late 1600s on one side of my grandmother's although the last name on that side doesn't sound pure Latvian but just and "s" added to end of last name likely heading back to Swedish (or perhaps German or Dutch) roots; probably quickly back to Germany on the other side of my grandmother; other than for the one trace way back the rest isn't really known past my great grandparents))

      k1a4a1 haplotype mitochondrial DNA

    5. part 2:
      The bizarre:
      Very bizarre is that the autosomal Geno 2.0 test gave my mom 2% Oceania (the rest of the percentages match closest to their Dutch population sample (which might perhaps be similar to what one from Latvia would be too?) Dutch is: 53% Northern European/30%Mediterranean/16%Southwest Asian; hers is: 51% Northern European/29%Mediterranean/17%Southwest Asian/2%Oceanian)!

      Not sure if the 2% component is supposed to imply some pretty wild tale where about 6 generations back a sailor on one of the early European explorations to say Tahiti (I could be wrong, but I believe that New Guineans had less favorable thoughts to outsiders at first and thus less likely early mixing???? whereas supposedly Tahitians were quite open??? so I mention Tahiti instead of New Guinea or other islands) married a pure Tahitian and then they both returned, or at the least, one of their children returned to Europe??

      That sounds both pretty darn cool and also a bit hard to believe since in this case it would have had to have happened so early on and with a very quick return back to Europe. It might be a bit easier to believe if it could have occurred more recently (but that would seem to conflict with likely family history or % Oceanian).

      I mean I do something like that did happen a very, very few times early on, but only a very, very, very few times, basically to the vanishing point.

      Napolean did invade the Baltic area and such I believe and I also believe some of his mean stayed behind and some of his soldiers might have been on South Pacific expeditions. The Dutch did hit up the New Guinea area very early on and one last name on the family tree could be Dutch (or German or Swedish). The English hit up Tahiti early on.

      I don't know the whole thing sounds both really cool and maybe far-fetched.

      Anyway I had not other explanation so that is what I was guessing at, as wild as it sounded, a few days ago when I heard the results.

      But I guess there is the other option:
      Perhaps the Geno 2.0 signals they test for Oceania might not actually have ended up all being strictly what they thought. Perhaps not all signals they look for are ones that came up only after their isolation, but instead a few markers they use for Oceania might actually be markers for some very ancient component that got mixed away and lost for the most part but well retained in Oceania and, for one odd reason or another, also got also retained in certain Baltic/Finnish/(and other?) sub-populations where there was some very long and extremely insular line going way back?

      Does anyone have any up to date thoughts? Is the latter theory considered likely now or something along it's lines?

      Also I spotted on one of those ancient human tests using one of the public domain testing admix tools where they also also tested tons of others from many countries that it seemed like people of heritage from a decent number of countries actually registered a bit of Oceania using whatever components they used to signify Oceania. Other than for Malaysians and near Malaysians (and Oceanians themselves) it seemed like most of the components were likely super small maybe .5% and less, although it was hard to tell since no hard #s just pixel graphs, some looked like they could be 1%, not sure any looked like 1.5-2%, but it was really hard to tell, maybe?

      Too bad we can't get results from grandparents, if the results popped up to 4-6% for one then I'd think that would imply the explorer theory! If it stayed 2% at most then I guess it would hint more at something along the latter explanations offered?

      Anonymouse of Anna, do you have any results from your parents or grandparents?

    6. Oops that last sentence had "Anonymous or Anna" somehow come out as "Anonymouse of Anna".

      Anyway, a third suggestion is that the 2% results might often be nothing, just noise.

      Geno 2.0 seems to think they are solid down to the 2% level though.

      Going back to the original hypothesis, that the 2% Oceanian signal is both recent and real:

      Interestingly, I just found a case where the director of Geno 2.0 discussed a mystery 2% Oceanian component with none other than Robert Ballard himself. He took it seriously and they guessed that it might have traced to a six generations prior mingling and noted that he had some Dutch relatives and all of the Dutch activity in the Melanesian region in the 1700s.

      So perhaps these signals are recent and real?

      And thinking about it more, sure we know the direct paternal line on my grandmother's side going way back, but so what? It's not like the 2% Oceania came from the mtDNA part of the test and, although I need to verify, I think we like only know about 4 of the 48 slots on my grandmothers side in the 5th and 6th generation era, i.e. not much really. (Also, Tahitians have a somewhat different origin than Melanesians so, just guessing, they might have come up as say 40-50% Oceanian per Geno 2.0 instead of 80-100% (and with a higher Asian component?) so you could perhaps get the 2% from 5 gens back Tahitian).

      Even among those who say they have a pretty detailed knowledge going back, I wonder if they really have all of the 64 and 32 slots of the direct 5th and 6th generations back filled in or maybe it's really only a half or even a quarter?

      It would also be interesting to know the statistics for how many with largely European heritage ended up testing to 2% type values for Oceanian (or Native American). My vague impression is that it's almost feeling like there might be suspiciously many, especially since only a small fraction of people have been tested and how relatively considerably rare this whole scenario I'd imagine should be, but I really don't have too much to base it on. I think I'll contact Geno 2.0 and inquire. It might make for an interesting quick little study for them.

  24. I took my DNA test with 23andme, and I find the results both intriguing and informative. The Relative Finder could be improved though as its frustrating to find people who are willing to share information. Most 23andmeers are there for Health and not Ancestry, which could change because of the new price. I also ran a Y test for my father through Ancestry a few years ago. Way back then I found the Haplogroup information fun, but never really got much else out of the experience. Their matching map showed results, but no one would reply to my messages. So, I wasn't able to connect with cousins. I also find it relatively frustrating that they wouldn't include older test results with the new upgrades. I spent good money for results and I feel like I got a bit cheated with Ancestry. With 23andme I didn't have to pay an extra penny to be included with their website upgrades. All the changes are there, and I have access. Better deal in my opinion. I've even bought a test for my father to take with 23andme, and will add additional family members in the future. I'd like to try Geno 2.0, but it's way too expensive. Especially when you're considering buying multiple tests.

  25. Thanks CecCe for this very informative post. I've been doing genealogy for over 35 yrs but hadn't taken a DNA test until just recently. I ordered the AncestryDNA test and just got the results:

    54% West African
    34% British Isles
    7% Scandinavian
    5% Uncertain

    My initial impressions are that the test missed some of the heritage I know I have from my research. For instance, there's no mention of my maternal German heritage (17th century immigrants) so that was a surprise.

    I confess I was also surprised at the Scandinavian finding which AncestryDNA attributes to Viking heritage. That could be from English roots that go back many centuries but you would think the test would pick up the more recent German heritage first.

    One thing I was trying to confirm or not is whether I have Native American ancestry as our oral family history suggests. But with the AncestryDNA, there's no report on negative findings - just that mysterious 5% of something uncertain.

    Isn't that a fairly hight percentage. Based on your earlier post about the percentages attributed to each ancestor, my 5% unspecified is almost a whole 2nd great grandparent!

    I wish I had read your post before I ordered AncestryDNA. So as you suggest, I'm going to follow up with a 23andme test and compare the findings.

    1. Hi Karen,
      Many people have been quite surprised by the Scandinavian percentage in their AncestryDNA test.
      As far as the 5% Unknown - that is fairly high, but I am glad that they aren't trying to force it to fit in a category if they really cannot identify it. 23andMe also categorizes some DNA as "Unspecified", however they appear to be doing a very good job of picking up small amounts of Native American.
      I will be interested to hear what you find, so please follow up after you receive your 23andMe results.
      Thanks for the comment,

  26. Hi Cece,

    I've never met my father, but a family friend said he was Jewish. I'm female. Which, if any, of these tests/companies could confirm my father's background for me? I don't have any male relatives that could do testing with me. Also, are any of these tests/companies able to do accurate BRCA testing? Finally, except for the BRCA testing, I really don't want to know if I'm suseptible to any dread diseases - I'm a worry wart! Is there any way to skip the medical testing with 23andMe? Thanks in advance for your insight!! Janet

    1. Hi Janet, Yes 23andMe will definitely confirm if you are part Jewish. You received 50% of your autosomal DNA from your father, so many of your matches will also be from his side. 23andMe does do some BRCA testing as well, but they don't do the full BRCA sequence since it is patented by Myriad Genetics. You can opt out of the health results there, but then you wouldn't be able to check your BRCA status. You could ask someone else to look for you though.

  27. Thanks for the great info Cece!

    Regarding 23andme's health testing, I've scoured their web site as well as the internet and I can't find specifics of their opt in/opt out policies and procedures. Is the health testing all or nothing? Or, for example, can a person opt in for the drug sensitivity testing but opt out of the rest of the health testing? Do they not run the tests you don't want or just keep the results from you? At what point in the whole process are you allowed to make these decisions? Thanks again! Janet

  28. Hi "Your Genetic Genealogist" ,I love your blog ,
    I had my test results with 23andme , looking to A.compisition I see most my dna come from Europe ,but also there is a asian/Native American than in Conservative view show 4.5 , in standard view show 6% and in especulative show 7.1%
    witch one it is more accurate the standard ,conservative or speculative?
    after professor MacDonald ,took a look to my raw data and he increase my native american to 8 % so what show I trust more the result of 23andme or the number that give me miter MacDonald, it he may be overestimating the native american percent ?. also I sent to him my father Data , and he show 14% native american ,
    here is his the result of MacDonald ,talking about my father and my 23andme raw Data ,
    Most likely fit is 74.5% (+- 3.3%) Europe (all Western Europe)
    and 14.1% (+- 0.2%) America (various subcontinents)
    and 11.4% (+- 3.3%) Mideast (various subcontinents)

    The following are possible population sets and their fractions,
    most likely at the top
    Spain= 0.751 Columbia= 0.140 Palestin= 0.110 or
    Spain= 0.751 Maya= 0.143 Palestin= 0.106 or
    Spain= 0.771 Columbia= 0.139 Egyptian= 0.090 or
    Spain= 0.769 Maya= 0.143 Egyptian= 0.088 or
    Spain= 0.686 Columbia= 0.140 Jewish= 0.174 or
    Spain= 0.770 Columbia= 0.139 Bedouin= 0.091 or
    Spain= 0.770 Maya= 0.142 Bedouin= 0.088 or
    Spain= 0.695 Maya= 0.143 Jewish= 0.162

    The Euro is either what the above says, or very far southern Italy
    (Which looks like Spain plus Mideast). The American looks
    North American. There is 0.7% sub-Saharan African which also
    is typical of either Spain or very far southern Italy. Note the 14% American ....
    23andme is notorious for getting this wrong.

    Most likely fit is 55.0% (+- 10.9%) Europe (all Western Europe)
    and 7.9% (+- 0.5%) America (various subcontinents)
    and 37.0% (+- 10.9%) Mideast (various subcontinents)

    The following are possible population sets and their fractions,
    most likely at the top
    Spain= 0.438 Columbia= 0.080 Jewish= 0.482 or
    Spain= 0.444 Maya= 0.082 Jewish= 0.474 or
    Spain= 0.657 Maya= 0.084 Palestin= 0.259 or
    Spain= 0.657 Columbia= 0.081 Palestin= 0.262 or
    Spain= 0.442 Karitian= 0.073 Jewish= 0.485 or
    Spain= 0.661 Pima= 0.082 Palestin= 0.257 or
    Spain= 0.430 Na-Dene= 0.089 Jewish= 0.482 or
    Spain= 0.451 Surui= 0.074 Jewish= 0.475 or
    Spain= 0.663 Surui= 0.075 Palestin= 0.262 or
    Spain= 0.659 Karitian= 0.074 Palestin= 0.267

    and this looks just the same as the provious one, except less American.
    Note that one does NOT expect exactly 50% as much American in the
    second as the first. This is because of “recombination” . One does,
    however, expect that, on average, a whole bunch of full siblings
    would have half as much as the sole American-bearing parent has.

    Note that the list of pictures is different ... some do not exist
    for Claudio.

    Doug McDonald

  29. Hi Cece , in 23andme result show my native american at 5% I sent my raw result to Doug Macdonald and he increase until 8 % , for what is more accurate what 23andme stimate or what MacDonald calculate ?
    I know that he always find more native on many users ,but am not sure if he is maybe overestimating the native american % .
    thank you

  30. Thanks for your work, it's really exciting
    My family is 100 % french, to be exact Norman.
    But, there's a tradition which says that one of my 3rd Great grandparent would be russian. It's why I decided to make analyses.
    In Ancestry finder I share a same segment with people from russia and eastern europe.
    So, AF confirms the oral tradition.
    Where I see a contradiction, in Ancestry composition, no eastern european .
    I also saw in global similarity map that residents of western Russia were included in the northern population.

    My results for the speculative version : 86,9 Northern European
    30, 6 nonspecific NE
    2,6 " " European
    My results: 32, 1 British Irish
    21 French German
    3,1 Scandinavian
    0,1 Finnish

    10,2 Southern European
    1,9 Iberian

    And it's just, where my chromosome is white that I share a segment with these people from Eastern Europe( on the first ), where it's nonspecific

  31. I am sorry , I precise, I have been tested at 23andme

  32. I want to do this, but have a question... since I cannot get a complete record being female, should I have my brother test as well to get the paternal haplogroup? I may have misunderstood, but am referring to the info here: https://www.23andme.com/ancestry/deep/#neanderthal

    All help appreciated!

  33. That's fine for someone that is one hundred percent pure white that already knows down to the exact city what countries in Ireland or Germany and they want the test to tell them something that they already know. 23andme maybe good at breaking down European countries but apparently they're not yet so good when it comes to black test takers getting these tiny false percents. 23andme consistently gives most African American test takers these very tiny "0.2% to 1.7% East Asian & Native American" results,when they have absolutely not one single trace of either Native American nor East Asian ancestry (if an African American REALLY has true Native American or East Asian ancestry they would get a much higher result),but other companies' admixture tests,like Familytreedna's Population Finder, don't give their African American test takers these false results. Why is this?

    1. I don't know the answer to your question, Reggie. You should post this question on the 23andMe community forum.
      23andMe does have an update coming for their AC that may improve any perceived deficiencies on this issue.

    2. Looks like you're wrong. They do tend to carry small amounts that are real. They were both enslaved together for a period of history

  34. My known etnic background is
    62.5 % Norwegian
    25.0 % Finnish
    12.5 % Finish or Saami
    Recent 23andMe AncestryComposition reads
    37.6 % Finnish
    24.5 % Scandinavian
    33.4 % Nonspecific North European
    4.0 % Nonspecific European
    0.1 % East Asian
    4.0 % Unassigned
    Do you think it is possible to gleen something
    about the Finnish/Saami ratio of my 12.5 %
    ”mystery genes” from the 23andMe results?

    1. Hi Arnljot,
      That is interesting that much of your known Norwegian was not assigned to Scandinavian DNA. As you can see, some of mine didn't either.
      The Finnish + Saami percentage looks to be almost exact though! I don't think any of the Norwegian bled into your Finnish percentage unless, of course, you have some Finnish ancestors in Norway. (Mine didn't.) Finnish DNA is quite distinct from that of the surrounding countries. So, that percentage is likely the 12.5% Saami + 25% Finnish. I'll bet that tiny bit of East Asian is from your Saami genes too.
      What else are you looking to learn from it?
      Thanks for commenting,

  35. My far too low % Scandinavian simply seems to stem from the low recall of only 0.36 (!!) for this group reported by 23andeME itself
    https://www.23andme.com/ancestry_composition_guide/. That alone remove/explains virtually all my Nonspecific North European.

    Most of my Norwegian family consists of inland valley famers that looks as arch Norwegians as they come. My chromosome map for each pair shows no Finnish ancestry for one of the chromosomes(my father). My X-chromosome is also more that half Finnish. Basically no problem here when the 23andMe recall is taken into account.

    Everyone in the family also seem to agree that both my grandmothers on my mother's side are 100% Finnish. The only uncertain grandparent is my mother's father. Here there has for decades been a sharp dispute within my mother's family concerning whether he is finish or saami.

    The big question for me is therefore how do the 23andMe Ancestry Composition score Saami in the presence of Finnish DNA?

    When I look at what happened to my arch Norwegian farmers it would not surprise me if the Saami suffered a similar fate in the 23andMe Ancestry Composition.

    I have corresponded via 23andMe with one lady with Saami ancestry who tells me that her cosine has a presumably 100% Saami grandmother that shows up in her Ancestry Composition only as Nonspecific North European. The ancestry of this lady is also reported as 7% Finish, but there are several possibilities for the origin of this Finnish DNA.

    Do you know of other examples of how 23andMe report the ancestry of persons with a known mixture of Saami and Finish? I am more that curious about this.

    1. I do not. The best place to ask about it is the 23andMe Community Forum. I find that I receive very interesting and insightful responses whenever I post inquiries there.

      I have a hard time believing that Saami would show up as Nonspecific Northern European since it is quite unique. Nonspecific NE usually implies that DNA is found throughout Northern Europe, which would not be the case with Saami DNA.

      Please let me know if you hear anything of interest. I am curious too!


    2. P.S. - I would think that your 4% Unassigned is more likely indicating the Saami DNA.

  36. I love your blog, CeCe. I had a DNA test done earlier in the year through Ancestry. Didn't do much research about different DNA companies at that time. Now, I'm getting ready to order a test for my ½ sister and am thinking, after reading your and some other blogs, that perhaps 23 & Me is the way to go. And I'll order one for myself as well, just to compare. Has anything happened in the past year (since most of the activity on this blog) that would change my mind about ordering a 23 & Me test?

  37. I love your blog, CeCe. It's been enlightening. I got Ancestry DNA tests for my husband and me earlier in the year. I'm preparing to order a test for my ½ sister for her birthday and decided to do some research on different DNA companies. It's very confusing but it sounds as if most people are saying 23 & Me is the best/most reliable. Most of the comments on this blog are from 2012 and 2013, have there been any changes since that time that would change your mind about recommending 23 & Me?

  38. This is a wonderful blog! I am wondering how you knew you have some Ashkenazic ancestry. The .4 percent figure seems so small! How many people/generations does that represent?

    1. Thanks Andrew.

      I know that the Ashkenazi Jewish prediction is correct because I have so many Jewish DNA cousins (who I share DNA with on my 7th chromosome). I have been able to map this segment of DNA back to my paternal great grandparents Blanche Purdy and Willard Moore so far. I don't know how far back the Jewish ancestor is. My sister's DNA is predicted to be over 1% Jewish. When you get down to those very small amounts it could come from anywhere between a 3rd great grandparent out to a 10th or maybe even further back.

      Thanks for your comment!

  39. If results are puzzling, be sure to look at history. I have always thought I was half Norwegian. Results show lots of cousins in Finland. As far as I knew, there were no Finns in my background - ever. But, it turns out that HISTORY shows that when Sweden ruled over Finland, it moved many Finns to border of Sweden and Norway to clear the forests. And guess what. The main city to which they moved was the exact same city my gre gre great grandparents were from. I learned that people even spoke Finnish there. So when those people came to US and said what country they were from, they naturally would say they were Norwegians because that is where their families had lived and intermarried for generations. DNA tells more of the history. My point being that just because your ancestors told you they were from a particular country, there may be more to the story. I also have seen Norwegians who wrote that they were Swedish in some census forms...because when THEY left, Sweden controlled Norway. So they lived in what later would become a Norwegian city.
    Think of the Germans who were invited to settle in Russia by Catherine the Great. They might tell their grandkids that they were Russian but they, in fact, belonged to a special isolated group of people called "Germans from Russia." Populations migrated all over and that is what makes this all so much fun.

  40. CeCe great blog and a fantastic review on various testing laboratories. I am very curious about knowing my ancestry. Both my parents were from South Asia. I am wondering which one of these testing laboratories are well equipped to unlock the genetic make up of South Asian populations. Appreciate any recommendations you can give me. Thanks

    1. Hi Curious,
      Thank you. This post is somewhat outdated, so I am glad you asked. I actually have a fair amount of experience with South Asian/Indian testers and have found that 23andMe has offered the most helpful results for that population group in my research. That said, none of the companies do a great job with Asian populations, so you may not get the answers you seek. It is also important to note that Family Tree DNA does have an India DNA Project, so that might be of interest to you as well. Best of luck with your research! Please check back and let me know how it goes after you test.

    2. CeCe, thanks very much for your prompt response even though, I posted on your outdated blog. I will try the two sources you mentioned.

      Thanks again. Keep up the good work

  41. CeCe, excellent blog and a fantastic review. Both my parents were born in South Asia. I am very curious to know about my genealogy. Which one of these laboratories you reviewed here is capable of analyzing South Asian genealogical make up? Appreciate your help on this. Thanks

  42. Who is currently more accurate when it comes to Finnish ancestry, Ancestry or Family Tree DNA?

    1. I find that 23andMe is actually the most accurate for Finnish DNA. FTDNA is next best in my experience Finnish. AncestryDNA includes some Russian reference samples in the Finnish/Northwest Russian panel which, in my opinion, throws it off a bit.

    2. Alright. In AncestryDNA my percentage for Finnish/Northwest Russian is 19%. In FTDNA, using the AncestryDNA raw data file, it's 9%. A great grandmother of mine is supposed to have been from Finland, possibly Sweden. The records are inconsistent and confusing.

    3. If you have no other known Finnish ancestor, then I would say your results have confirmed this oral history. Are we cousins? ;-)

  43. Firstly, I would like to thank you for your most current reviews of the various DNA packages available ( at a reasonable cost I might add). I read in this blog based on your critique of the NatGeo testing that you feel that Spainish Portuguese aspects of the test did not accurately reflect your DNA model. Interestingly, in the other tests you do have British Irish percentages which for whatever reason assumes to group them together and that based on my own laypersons experience is a fault in that presentation of the results. The NatGeo testing presentation seems to take into consideration human migration patterns. The Irish aspect of your genealogical record would in fact have a Portuguese Spainish component in it as research based on human migration and cultural history shows that the Celtic component is partly sourced from the northern regions of the Iberian peninsula. Did you know there is town/soccer club known as Celta Vigo? That isn't coincidence, but a historical sliver of that Iberian peninsula area's Celtic origins. There are supposedly even remnants of that ancient dialect in the modern local Spainish/basque. I think the key flaw to all this DNA ancestry testing lies in the various approaches in interpretation. Refinement of those interpretations should include a more through synthesis based on a combination of the genetic and most current historical research models from the anthropological current. Separation of the percentages based on ethnicity would also go a long way to better tell the DNA story. From what I have seen based on your analysis of some of these tests, the major flaw is the presentation of the results and not the results themselves.

  44. Hi!
    Coming a bit late into the diskussion... you talk as if finns, saamis and Scandinavians were living on different planets. I live in the Botnia Bay area and are happy to inform you that all groups here have live side by side för many, many generations and there were no laws against intermarriges. A lot of movements and new settlements all over the wast area, all groups represented and coming together . So if there is a blend in your "pure" heritage, rejoice, your ancestors were not to inbred to survive.

    1. Thank you for your perspective, Margareta. Finnish DNA is not very similar to that of their Scandinavian neighbors which demonstrates that there was not much intermarriage between them over the last few hundred years. It doesn't make much sense to me as to why this would be, but that is what the DNA is telling us. Maybe your region is different than where my ancestors came from in that regard. Thank you.

  45. Hi there :-)
    which service do you think might be best if I am trying to trace Sephardi Jewish ancestry? Oral history in my family is fleeing Spain and Portugal due to the Inquisition.
    many thanks, Rivka

  46. How accurate is your original post considering it's well over 4 years old? For instance, AncestryDNA autosomal wasn't even offered back then. Also, the Genographic Project has radically changed since then, and their results are more complex. Do you have an updated comparison post?

    1. No, I do not. I have been thinking about doing one, but haven't had time. Obviously, anyone who reads the post needs to take into account the date on it. AncestryDNA was indeed offered in 2012. Otherwise, they would not have been included here with screen shots. I was an early beta tester, but by the time of this post, they were testing lots of members. With that said, my preferences and opinions that I laid out in this post have not changed much in the ensuing 4 years. I still find 23andMe to be the most accurate and informative for admixture. Next would be AncestryDNA because they have improved a lot over the years, but are still not as informative due to the lack of the chromosome view offered by only 23andMe. I continue to have issues with FTDNA's admixture, now called MyOrigins and I have not tested on the new Genographic platform.