This is a public Forum  public


    Lynn Recker
    Trading a Scandanavian problem for a Great Britain problem?
    Question posted November 1, 2013 by Lynn ReckerLevel 1  , tagged AncestryDNA, DNA, DNA Results
    448 Views, 2 Comments
    Trading a Scandanavian problem for a Great Britain problem?
    Sampling of Europe shows a relative shortfall in the Europe West area compared to other areas of Europe

    First, I want to say that I think AncestryDNA has made some great strides in determining one's ethnicities when compared to the original determinations.  I am also happy to hear that AncestryDNA considers that the ethnicity estimates are a work in progress - because I do believe that AncestryDNA has traded what was perceived to be a "Scandanavian Problem" for a newer "Great Britain Problem."

    AncestryDNA received a lot of outcry (and perhaps rightly so) about persons with too much Scandanavian ethnicity. That was certainly one reason why the new ethnicity determinations were instituted.  Reading the complaints at the time, it seemed that the problem was perceived by persons from AncestryDNA to primarily have to do with those who are ethnically British.  However, there were those of us whose ancestry derives from continental Europe who also had a "Scandanavian Problem," which has not been fully addressed.   

    I have come to think that AncestryDNA has not considered fully the idea that some (many?) of us with Belgian, Netherlands or NW German ancestry also had the "Scandanavian Problem."   My "v.1" ethnicity included a rating of 76% Scandanavian.  My "v.2" ethnicity now has me as 45% English, 25% Europe West, and only 6% Scandanavian.  It seems that  prior 76% has been split among the three regions, with an emphasis on Great Britain.  However, I haven't one "Great Britain" ancestor just as I have no known Scandanavian ancestors (since abt 1670).  A quarter of my ancestry derives from northwestern Germany (Niedersachsen), and another quarter is Belgian, split equally between ancestors from Brabant-Wallon and Luxembourg.    Yes, I realize that the Great Britain DNA signature is detected in small portions in many neighboring regions including continental Europe.  However, I would ask if that is truly what should be called a "Great Britain DNA signature."   Instead I wonder if a listing of some of the DNA signatures as "nonspecific northern European" at this time would be a more fair determination - especially as there seems to be a bias against the Europe West ethnicity in your sampling. 

    Here is what I found (from the AncestryDNA White Paper on Ethnicity) when I checked into the relative numbers of those Ancestry has in their reference sample when comparing their Ethnicity Regions of: Great Britain, Scandanavia, East Europe, & West Europe.

    The total population of these four ethnic regions is: 450 million (true not everyone in Europe these days is "European" - but the relative numbers are surely meaningful). The total reference sample for these four ethnic regions is: 941 persons. The following listing shows:

    Ref # - the reference panel number for each ethnic region;
    Ref % - the reference panel number as a percentage of the total sample for the four ethnic regions;
    Pop - the population (in millions) in the core of the ethnic region;
    Pop % - the population as a percentage of the population for the four ethnic regions; and
    % Ratio - the ratio of the percentage of the reference panel number to the percentage of the population.

    If the ratio of those percentages (% Ratio) is less than 1 (unity) then the representation of that region's DNA count in the sample may be below par and under-represented in calculations. (Rounding may not allow percentages to add up to exactly 100%).  Now I have made a big assumption here, and I don't doubt that someone at AncestryDNA may be able to refute it.   Here are the numbers I determined.

    Ref # -- Ref % -- Pop -- Pop % -- % Ratio
     111 --- 11.8% --- 63 --- 14% ----- 0.84

    Ref # -- Ref % -- Pop -- Pop % -- % Ratio
     232 --- 24.6% --- 20 --- 4.5% ----- 5.47

    Ref # -- Ref % -- Pop -- Pop % -- % Ratio
     432 --- 45.9% --- 165 -- 36.7% --- 1.25

    Ref # -- Ref % -- Pop -- Pop % -- % Ratio
     166 --- 17.6% --- 202 --- 44.9% --- 0.39

    Note that the Scandanavian Reference Panel is grossly over-represented in the Ancestry pool. Is this the reason for the prior "Scandanavian Problem" which Ancestry had in their first determination of ethnicity?  Europe East is over-represented as well, but close to unity.  Great Britain, though below unity, is well represented relative to that of Europe West, which is the source (Saxons) of much of that region's ethnic heritage. Which brings us to Europe West which is very poorly represented at a ratio of only 0.39.

    From this I conclude that it is quite probable that the Saxon element in Great Britain is of a greater representation than in its homeland of Germany. Further it may be probable that the Celtic DNA elements of the low countries, the Netherlands and Belgium, do not have the representation needed to discern a differnce between that of Europe West and Great Britain.  However, only AncestryDNA which knows the make-up of their representative samples can answer the questions raised here. 

    Maybe what this means is that my German Saxon and Belgian Brabanters are a representative remnant population of those Saxons and Celts which eventually settled in Great Britain. If so, I would rather that those British with similar DNA be identified ethnically as from Europe West, which would be indicative of their deeper ancestry. That would be more acceptable than to have my (and other's) ancestors, who had no association with the British Isles, be called ethnically "Great Britain."

    Moreover, I simply believe it would be more honest in listing certain DNA as "nonspecific northern [northwestern?] European" at this point in time.  When AncestryDNA applies their shoehorn to the data they are unfortunately attempting to put some people into places that they may not belong. My people are all continental European and most go back to the 1600s (I do not appear to have a disconnect as I have some connections by DNA on both sides of the family, at least one back 8 generations). On the off chance some English-person(s) made their way onto the continent and into my gene-pool, I cannot see how they would have overwhelmed my DNA creating a Great Britain signature of up to 45%.

    Thank you for the forum to air my views.



    • Anna Swayne

      Thanks for your post, Lynn. We appreciate your comments, and understand you have put a lot of thought into the complexity of your DNA results. We have had several members of the DNA team review your post. We hope you will find the following points helpful in understanding your ethnicity estimation and how it was calculated.

      • Certain regions have had more "mixing" of DNA due to human history, invasions, and migrations, and are harder to tease apart.  Europe West and Great Britain are two of these regions.  The following link is to a blog post where we explain the science behind the ethnicity estimation.
      • Differing sample sizes in our reference panel could indeed have effects on our estimates of ethnicity.  To this end, AncestryDNA is actively working to analyze more DNA samples from individuals around the world to further refine our ethnicity estimates.
        • From the latest ethnicity update we want to be as transparent as possible, so that any individual can see how they compare to the population data of any one of the regions. As time goes on you may see your current ethnicity estimate change based on more research for specific regions.
      • We appreciate your feedback on renaming the region to "nonspecific Northern European." However, 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.  While the "Great Britain" genetic ethnicity is not just found in people from Great Britain, individuals from this region form the basis for the ethnicity estimate.
      • In addition to our sample collection, however, further developments in our algorithms can also help us to reach better refinements to ethnicity estimation.  This may allow us to reach finer resolution in some of these more difficult areas in the near future.  
      • Our ethnicity estimates are a fun part of our genetic story and remember it’s just that-an estimate- and comes with a likely range for that estimate. Check out your estimated ranges for each region, you may find insight in the breakdown ranges that estimate what you were expecting.

      We will continue to work on developing the science and technology behind the AncestryDNA test and provide you the tools to discover your genetic story.

      Thanks again for your comments and feedback.


      AncestryDNA TEAM


      • Lynn Recker

        Thank you very much for your answer. 

        I wasn't thinking of renaming Great Britain specifically into "nonspecific Northern European."  Instead, I would think that there may be a signature that could be found for Great Britain that might be differentiated from a more continental one.  Yes, such a differentiation would be rather difficult.   And the only way it could be done is to increase the representative sample of indigenous peoples from the low-country areas.  And yes, it would necessarily mean that some who have British ancestry might be classified as European.  But at least you are pointing to their deep ethnicity; whereas with persons such as myself, AncestryDNA is misidentifying my genetic heritage.

        I'll look forward to DNA Ethnicity v.3 in the future.

        Thanks again, Lynn