Admixture: not soup yet

And maybe never

The Legal Genealogist wants everyone who’s ever taken a DNA test that provides a set of ethnicity or admixture estimate percentages to repeat after me:

“It’s not soup yet.”1

And it may never be.

So… while I was on the road (and coping with the now-dry-but-once-flooded basement), Family Tree DNA released its new myOrigins ethnicity estimates, a major upgrade from the old Population Finder.

myorigins2It’s a major step forward in our quest to gain information about our distant ancestors — where they lived, where they came from. It’s as good as it comes given the limits of science. It applies the very best of a set of statistical algorithms to what science knows to come up with a formula that tells us what it can about our origins.

And the very best of what it can tell us is still a guess.

Now you may be sitting there thinking, “but they don’t teach this in law school. How can she know?”

I know because I’ve spent a fortune testing relatives who’ve been kind enough to let me manage their results. And I can see what happens when, for example, you apply the very best of a set of statistical algorithms to what science knows to come up with a formula — and then apply that formula to people whose results should be pretty much identical.

People like my mother’s full blood siblings.

Four of them have tested. They all share the same mother, same father, so same grandparents and great grandparents on back into the mists of time. Their DNA establishes that fact beyond any question. No non-paternal events here, no undocumented adoptions.

So if all four have exactly the same parents, and grandparents, and great grandparents, and on back through the generations, they ought to have the same genetic admixture, right? Ethnically, looking back at their heritage 500 or 1,000 or 10,000 years ago, they can’t be different.

Except they are, at least according to this very best of statistical algorithms. Take a look at the chart below:


So… the population group labeled European Northlands “centers on the people of Scandinavia.”2 Does this family have any Scandinavian ancestry? Three of the four show traces, the fourth shows none at all.

Well, then, what about the North Circumpolar group — a group that “began around the arctic as hunter-gatherer peoples … (and) stretches from Lappland east to Greenland”? Three of the four show traces of that, the fourth shows none at all.

What about the North Mediterranean Basin population? A “distinct European cluster… situated in the southwest of Europe from Spain to Greece.”3 Three of the four show traces of this group, the fourth shows none at all.

We might be tempted to say yes to all three of those cases, because three out of four — 75% — of the siblings tested do show traces.

But what about the Trans-Ural Peneplain? The “dominant group between the tundra and the steppe in Eurasia’s northwest … from the area where the North European Coastal Plain joins the forests of Central Siberia.”4 One has it, three don’t. Eastern Afroasiatic? A group that “developed in the Persian Gulf north toward the Zagros Mountains.”5 One has it, three don’t.

How can this be?

We have to keep in mind what these admixture tests do: they take the DNA of living people — us, the test takers — and they compare it to the DNA of other living people — people whose parents and grandparents and, sometimes, even great grandparents all come from one geographic area. Then they try to extrapolate backwards into time. Nobody is out there running around, digging up 500- or 1,000-year-old bones, extracting DNA for us to compare our own DNA to.

So coming up with these percentages in these tests requires this fundamental assumption: that the DNA of the reference populations — those groups whose parents, grandparents, great grandparents and more all come from the same area — is likely to reflect what we might see if we could test the DNA of people who lived in that area hundreds and thousands of years ago.

In other words, these percentages are:

• estimates,

• estimates based on comparisons not to actual historical populations but rather to small groups of people living today, and

• estimates based purely on the statistical odds that those small groups tell us something meaningful about past populations.

These limitations are true of all of the testing companies, not just Family Tree DNA. I’ve tested with them all, and my own results are — literally and figuratively — all over the map. I’m German with some companies, not German at all with another. Largely Scandinavian with one, only slightly Scandinavian with the others.

It isn’t soup yet.

And because we aren’t about to go digging up those old bones, it may never be soup.

DNA testing for genealogical purposes is a wonderful tool. But people get disappointed when they see these percentages and they don’t match up to their own paper trail and don’t match up from company to company. And when they get disappointed, they may lose interest in genealogy or in DNA testing. And when they lose interest, we lose out on the paper trail information they might add to our mix.

Bottom line: We need to educate our friends and our families, our DNA cousins, to the limits of what these percentages can show — and to show them all the other things DNA testing really can help with.

Because friends don’t let friends do DNA testing only to get these percentages.


  1. For those too young to remember the reference, the Lipton Soup Company had a string of ads in the late 1960s and early 1970s. The mother in the ad would begin preparing Lipton soup, a child would ask, over and over, “is it soup yet?” and the mother would answer “not yet” over and over until finally she’d say, “It’s soup!” So “not soup yet” means “not finished” or “not ready.”
  2. Population Clusters in myOrigins,” Family Tree DNA Learning Center ( : accessed 17 May 2014).
  3. Ibid.
  4. Ibid.
  5. Ibid.
Print Friendly
This entry was posted in DNA. Bookmark the permalink.

42 Responses to Admixture: not soup yet

  1. Tim Campbell says:

    Thanks, Judy! I too have tested with Family Tree DNA and I too was amused when I saw myOrigins. The 95% European didn’t change (didn’t surprise me since I’m so white I’m OMG!) but the other 5% skipped a whole continent.

    I did a statistical analysis of my own and realized that Family Tree DNA only has a fraction of 1% of the world population to compare results (same with every other testing company). Genetic DNA isn’t there yet and won’t be until a few million more people get tested.

    I guess I better get out and promote DNA testing.

    • Judy G. Russell says:

      If I can offer a minor caveat… Genetic DNA is there. It’s just this one little part of it, the admixture part, that isn’t — and may never be there in the kind of detail we all want. (I don’t know about you, but forget just telling me what country my DNA comes from — I want every last jot and tittle of my DNA to have a name tag!) We absolutely can — and do — use DNA for all kinds of research questions right now. But answering the question of exactly where in the world our ancestors came from in more detail than the broad continental basis? Not now, maybe never.

  2. Peggy says:

    If one of the siblings had an estimate of greater than 25% of one of the reference populations and each of the other siblings did not at least have a trace of that same reference population then I think the results would be suspect. But as the difference between the siblings is all in those estimates of reference populations below 10%, it really just looks like the random shuffling of the deck that is DNA recombination. But hey, it’s been a long time since college statistics so perhaps someone more up-to-speed on statistics could comment?

    • Judy G. Russell says:

      I’m not a statistician, and don’t play one even in makebelieve life! — but there’s an awfully wide gap to my way of thinking between 69% European coastal plain and 39%!

      • Drew Smith says:

        For my ethnicity estimates on AncestryDNA, my Great Britain percentage is given as 30%, with a range of 0%-56%. The 30% represents the *average* of each of 40 different segments tested for a Great Britain ethnicity (apparently at least one of my segments showed no Great Britain ancestry, and at least one showed 56%).

        I have no idea how many segments Family Tree DNA tests for, but if they use a similar procedure, it would probably be more interesting to see what *range* they predict for each geographic area, and then to see if the ranges overlap between siblings.

        • Judy G. Russell says:

          I would love to see how each segment tests out, Drew! That’s a piece of information that might actually be useful for genealogy!

        • George P. Farris says:

          Drew, as with most things to do with DNA I feel clueless. But, since two great minds such as Spencer Wells and Judy Russell both express opinions that it’s a valuable tool I demur to their intellects.

          However, I am not grasping the percentage, range, or those 40 segments of Great Britain ethnicity that Ancestry DNA uses. Can you, Ancestry, or Judy clarify that a bit?

  3. Pingback: DNA As A Genealogy Tool – Not Ready for Prime Time? | Andrew J Morris

  4. Jay says:

    I’m still new to DNA testing, but I’m not impressed with FTDNA’s latest ‘myorigins’. I am a multiracial person and know most of my family history, (African, Native,-both of these are recent, G-G grandparents were African,- East European, French, Swedish, Irish, Scottish, East Asian, and West Asian),and FTDNA’s results didn’t pick up on anything but the Swedish and Irish, (aka European Coastal and European Northlands. I’ve tested with other companies and they showed some of the other racial components.

  5. Robert Hobdell says:

    What you say was pretty clear to me after reading’s explanation of how their DNA Ethnicity Estimate is generated. The problem is you do not get to access this information before you buy the DNA test (as far as I know). So thanks very much for enlightening potential customers about this. In my case the Ethnicity Estimate did not show me anything I did not already have a good idea of. My only surprise is I’m more Irish than expected. For those who are truly clueless about their origins an Ethnicity Estimate may point where to go looking. I think for most folks it is a waste of time.

    • Judy G. Russell says:

      I can’t stress enough that ONLY the ethnicity part is at issue here. There is so much we can all learn from autosomal testing — it’s just that these percentages aren’t really a good reason to test in the first place. So anyone interesting in genealogy SHOULD do autosomal testing, but for genealogy — for finding cousins and getting access to those parts of our family story that only those cousins have. And if we ever get decent percentage to go along with the autosomal cousins we find, well, that’ll be a net plus.

  6. Barbara says:

    I am new to the DNA testing. Just received my results and have been avidly reading and educating. I think what needs to be pointed out when testing siblings is the randomness of that 50% you receive from each parent. Just because you share the same parents does not mean you inherited the same genes. Each of us is a unique genetic combination. I am looking forward to seeing my brother’s results. We look nothing alike so I don’t need a DNA test to tell me we inherited different versions of those genes. Comparing our DNA results using the admixture testing will be very interesting and I will not be surprised if our percentages are way off.

    • Judy G. Russell says:

      Your admixture actually should be fairly close… but it’ll be interesting to see if it is!

      • Walker Hall says:

        I agree with Barbara. On AVERAGE, we get 25% of our DNA from each grandparent. But that’s an average. The possible range is 0%-50% from each grandparent. YOU may have inherited 10% Paternal GFather, 40% Paternal GMother; 10% Maternal GFather, 40% Maternal GMother. Now, assume the two grandparents you got a heavy mixture of happened to both be made up of a certain population type that neither of the other two grandparents had. Are you seeing the possibilities? If your sibling gets an opposite disbursement — or even just a more even disbursement — then what showed up STRONG for you might show up extremely weak in your sibling; or what showed up as moderately decent amounts for you may show zero for your sibling.

    • Walker Hall says:

      Did you get the results yet? I’d be curious to know. :)

  7. Lynn David says:

    AncestryDNA had it’s Scandanavian ethnicity so ramped up that people got mad at them. So they came up with v2 which still has its problems. Some I have pointed out here:

    Anyone should be taking these estimates with a teaspoon of salt.

  8. Pingback: My Native American DNA. Say what?

  9. Michel Guévremont says:

    I have requested Combo Chromo 2 tests from BritainsDNA and I a waiting for my results later this month or early Jan/15. I am confirmed Y-HP I1*(Ultra Norse Viking descent of either norwegian or Danish origin) (at least 5,000 to 8,000 Yrs old) and mtDNA HP H*. The star indicate undetermined. I

  10. j s says:

    I think you are sure of the four siblings parents, and perhaps grandparents, but how can you be sure of the parentage farther back? Would that not show up in the dna of the siblings? The only parentage we can be sure of is the mother to child line…

    • Judy G. Russell says:

      Judith, the point here is that — no matter what the distant parentage of the siblings turns out to be — all four of them have exactly the same parents and therefore the same grandparents, great grandparents, great great grandparents and so on. Presumably, an ethnicity test should record the same ethnicity for all of them. It doesn’t. It can’t. That’s why it’s at best an estimate.

      • Walker Hall says:

        That’s just the thing though. We are NOT the same ethnicity as our siblings — not identically; not when it comes to genetics. Sure, siblings have identical ethnicity in the genealogical sense, which basically assigns the same evenly-divided percentage to every ancestor. But genetically, siblings are not (exactly) the same ethnicity. On average, many families will have siblings with very similar ethnicity (but never identical!), and likewise statistics tell us there will be some families whose siblings have fairly different ethnicity. Obviously, if MOST families start showing up as having siblings with massively different ethnicity, there is a problem with the algorithm or something. But statistically, there will be a whole range of outcomes possible.

      • Walker Hall says:

        EDIT: Maybe the issue here is the definition of ethnicity? It sounds to me that you are thinking of it in terms of purely genealogical descent. However, we are clearly discussing DNA tests, so in this context, ethnicity must mean “what part of the world did your DNA come from” and not “what part of the world did your genealogically ancestors come from.” Since each sibling gets a different percentage of the grandparents DNA, they will get different percentages of those ethnicities, and hence siblings will never be perfectly identical matches ethnically unless they got the EXACT same percentage from each grandparent that their sibling did (I.E. 23.4532947562623758573…% )

      • Walker Hall says:

        PS: Let me add that I am loving your articles, in case you haven’t noticed from all my comments. haha And please take my comments as me thinking out loud, and asking for your input, confirmation or otherwise. I probably should phrase more of my “statements” as questions since I don’t mean to imply I am 100% sure but am rather simply trying to make sense of the issues. :)

        • Judy G. Russell says:

          As you note, defining ethnicity is at the heart of this, Walker. And the whole point here is that even if you could assign those very detailed percentages down to the one-kazillionth of a percent, we still wouldn’t know for sure that the admixture assignments were right. No matter what we do, we’re still comparing living people (you and me and our siblings who test) against other living people — and therein lies the rub.

  11. Linda Strickland says:

    I did the Ancestry DNA test, uploaded my raw data to FTDNA, and uploaded the date to Gedmatch. I just sent in a sample to 23 and Me. My reasons for testing and uploading are simple…. I needed ideas for which countries to search in to go further back in my family tree. In the results of Ancestry I found a first cousin that I never knew I had. He had been given up for adoption a couple of years before I was born. I am thrilled to find more and more cousins both recent and distant. I was not surprised by most of my Ethnic results except for having 26 % Italy/Greece region on Ancestry. I still have not found a paper link to any Italian person, but I have lots of fresh ideas on where to search because of DNA testing.

  12. Taz Gardner says:

    As someone with relatively rare comorbid conditions, one of which is currently being studied to eradicate it, I find it incomprehensible that so many people are paying to basically give up all rights to confidentiality concerning their medical status. As to determining ethnicity, I can’t help but wonder why, given the fact that it seems to add to xenophobia. I’m delighted to know that more people are diversifying when choosing partners with whom they want to procreate, and long for the day when we evolve enough to use this information strictly to develop medical treatments that are tailored for the ailing individual via their DNA. As long as all our DNA shows we are the same species, that seems sufficient to satisfy my ancestral curiosity.

    • Judy G. Russell says:

      I find it incomprehensible that anybody would think having a genetic genealogy test would “give up all rights to confidentiality concerning their medical status.” As the song says, “it ain’t necessarily so…”

  13. Chris Renk-Keese says:

    Hi Judy, I hope this site is still active. My question is how would I determine (approximately) my MATERNAL grandmother’s heritage/ethnicity? Which site, which DNA test, etc. My mother is 83 years old and the only living member of her immediate family. Both her parents are deceased. After my maternal grandmother’s passing, my grandfather shared her secret with all their children. My grandma was VERY ashamed that she was left on an English preacher’s doorstep in Chicago, in the1800′s. She made my grandfather promise to never share her secret with their children until after her death. Therefore, my mom’s 8 maternal cousins are not really biological cousins after all.

    I know my mom would LOVE to find out her ethnicity/heritage before she passes. I would LOVE to give her as close as possible an answer.

    How do you recommend I go about this? I’d obviously get her tested for mDNA, but if I tested and had my maternal cousins tested, would this give more accurate results?


    Christine Keese

    • Judy G. Russell says:

      You’re better off testing your mother, not you, when it comes to ethnicity. Yours will be mixed — your mother’s and your father’s. Hers will be closer to the people she wants to know about. And keeping in mind that none of the ethnicity estimates is better than just that — an estimate, and not set in stone — the general view is that 23andMe has the best ethnicity estimates right now particularly for those of European ancestry.

  14. A simple statistical analysis of the results for the 4 siblings gives:

    E. Coastal Plain: Average 53% Standard Error ± 7.5%
    E. Coastal Islands: Average 33% Standard Error ± 3.4%
    North Med. Basin Average 4% Standard Error ± 2%
    E. Northlands: Average 5% Standard Error ± 2%
    N. Ciercumpolar: Average 4% Standard Error ± 2%
    Trans-Ural Peneplain: Average 0.8% Standard Error ± 0.8%
    E. Afroasiatic: Average 0.5% Standard Error ± 0.5%

    Each individual sibling has a random sample of 50% of the parents’ DNA, but the average of 4 siblings’ results represents a sampling of 15/16 of the parents’ DNA, so gives a much more reliable estimate of the % values in each category.
    You can see from the Standard Error of the Average values that the uncertainties in the percentages in each category are fairly large.

    • Judy G. Russell says:

      >> the uncertainties in the percentages in each category are fairly large. <<

      To put it mildly. Thanks for the input.

      • Since the DNA sample from four siblings represents 15/16 (93.75%) of the parents’ DNA, the uncertainties in my earlier post are indicative of the uncertainties in the DNA test results.

        • Judy G. Russell says:

          And that’s been precisely my point all along: this isn’t soup yet, and it may never be soup (at least not until or unless we can factor in a broad enough sampling of ancient DNA).

  15. Anne Hart says:

    Happy most of my DNA is Greek. Always knew I love stuffed grape leaves and tomato salad with olives.

  16. E. Collins says:

    Hi Judy,
    I really have enjoyed your blogs and decided to post since you seem really good at replying to all comments and doing so in a timely manner. I plan to follow your advise and get autochoromsomal testing with AncestryDNA. My question is about which offers mtDNA and yDNA testing to determine your african ancestry to the continent and the ethnic group. The testing is pricey and I just wanted to know how accurate this could be and is it worth the price ($250)? Or should I just try to find out through research by matches to others family trees and through testing with other autochromosomal DNA compaanies (i.e. Family Tree and 23andme?)
    Thank you for yor help. All the time and effort you put into answering everyones questions is greatly appreciated.

    • Judy G. Russell says:

      (a) The DNA test other than YDNA and mtDNA is autosomal, just in case you need the exact name for any reason.

      (b) African Ancestry as a testing company does only YDNA and mtDNA testing, and no autosomal DNA testing. You may get to drill down a little better to African ancestors in the direct lines (your father’s father’s father etc. if you’re male or your mother’s mother’s mother etc. for both men and women), but you get no information about any other part of your ancestry — and that leaves out a LOT of people who contributed to who you are genetically. Once you get those results from African Ancestry, they can’t be used for US-based genealogical purposes: you can’t join a surname project, for example. So overall they’re not as useful for genealogy as testing with the major genetic genealogy companies.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>