What Dutch DNA looks like

There’s a meme going around on Facebook, started by J Paul Hawthorne, where people share their pedigree chart that shows the ethnicities rather than the names. In this article I am going to compare my ethnicity based on my paper trail to the ethnicity predictions by the DNA companies.

Ethnicity pedigree chart

Pedigree chart with ethnicity's: about 95% Dutch, a little German and some white (unknown) blocks

Pedigree chart of Yvette Hoitink with ethnicity

Here is my pedigree chart, showing the countries of birth of ten generations of my ancestors. As you can see from all of the orange, I am quite Dutch.

The blue lines represent people born in Germany. Most of my German ancestors are from Suderwick, just across the border from the Achterhoek area where the rest of my paternal ancestors came from.

I have some unknown ancestors, including three unknown fathers of illegitimate children on my mom’s side, which show as blanks.

My father’s family did not move around. All 2,000+ ancestors of his that I have traced so far were born within a few miles of the house where he was born in Winterswijk. Based on naming patterns and archaeology, it is known that many farms in this area have been around since about 800 AD. That may well have been the last period when large groups of new people settled there.

My mom’s family is a bit more mobile, with ancestors from Noord-Brabant and Zeeland. On her Zeeland lines, she has some Huguenot ancestors from northern France. On her Brabant lines, she has some bastards of nobles, which tie into the royal families of medieval Europe, but these are all at least fifteen generations back.

All in all, I think I’m pretty Dutch!

Admixture predictions by DNA companies

So what do the DNA testing companies make of this? I had autosomal tests done by three different companies, who all provide their ethnicity predictions. Here is how they read my DNA.

Ancestry

Ethnicity estimate by Ancestry

Ancestry’s ethnicity estimate is spot on, showing I’m 100% European, and 94% from Western Europe. The trace regions could be explained by the medieval noble lines, although they are so low they could easily be noise.

23andMe

23andMeEthnicity

23andMe has three levels of predictions: conservative, standard and speculative. All of them have me as predominantly Northwestern European, which matches my paper trail.

At the speculative level, 23andMe tries to assign even DNA it isn’t sure about. French & German, a region which includes the Netherlands, comes up as the highest. Even combined with the unassigned “Broadly Northwestern European,” that only accounts for less than two-thirds of my DNA. The 28.2% British & Irish is way too high, as is the 6.5% Scandinavian. So not bad, but not particularly impressive either.

FamilyTreeDNA

Ethnicity predictions FamilyTreeDNA

Boy, did FamilyTreeDNA miss the mark on that one!

None of the regions cover the eastern part of the Netherlands, where my father’s family is from. When I first saw these results, I thought FamilyTreeDNA did not have a reference population for the area around the Netherlands, but I later had other family members tested and they had “Western and Central European” which covered the Netherlands.

It’s really unbelievable how FamilyTree failed to pick up even one percent of Western and Central European DNA, where my estimate based on my paper trail shows it should be close to 100%.

Paper trail versus DNA

In the past, when I’ve used my results to warn people about the unreliability of these ethnicity predictions, some people tried to convince me that my paper trail must be incorrect. However, I do not think that is the case:

  • I match all the right people. If there was a recent non-paternal event, I would not expect to match all the fourth, fifth and further cousins that I do.
  • Most of my ancestors lived in small villages. Even if I made a mistake somewhere and got the wrong parents, chances are that the right parents were from the same village. That should not affect my ethnicity predictions, as those people weren’t immigrants but usually lived there for many generations.
  • My paternal uncle and mother tested as well. Their ethnicity predictions are quite different from mine. For example, FamilyTreeDNA shows significant Western and Central European DNA for both of them. DNA proves they are definitely my mother and uncle, so if they are Western and Central European, so am I.
  • The predictions of these three companies are quite different from each other. This in itself shows we should not put too much faith in them and that the prediction models are in its infancy.

To me, these results confirm what Judy Russell was saying about ethnicity predictions two years ago: “it’s not soup yet.”

About Yvette Hoitink

Yvette Hoitink is a professional genealogist in the Netherlands. She has been doing genealogy for almost 25 years. Her expertise is helping people from across the world find their ancestors in the Netherlands. Read about Yvette's professional genealogy services.

Comments

  1. Evelien van Niekerken says:

    Erg leuk om te lezen! Ik heb mezelf ook via 23andme en FTDNA laten testen. Hoewel het niet echt met elkaar overeenkomt vind ik het wel leuk. 23andme had mij ook een grote portie British toegekend terwijl mijn achtergrond toch echt vooral Nederlands/Duits is. Qua matches is het, denk ik, veelbelovend. Veel Amerikanen testen met Ancestry waardoor veel geadopteerden via deze weg hun geboortefamilies vinden. Kortom: fascinerend!

  2. Donald W. Blankman says:

    Hoi Yvette, for years I was content in the belief I had found the earliest occurrence of my surname line in The Netherlands. My most distant direct male ancestor seemed to be Pieter Blanckman, later Blankman, who lived in Oudesluis, de Zijpe, Noord Holland, ca. 1565. A few years ago I had my Y-Chromosome sequenced and learned that I belonged to Y Haplogroup N1c1, which is very rare in The Netherlands. Consequently, I now seems likely the my ancestor Pieter’s male ancestors were most likely from Finland, where 61% of the male population belong to that Haplogroup. Almost all of my matches are Finns. There is only one Dutchman in FTDNA’s Netherlands group who matches me, unfortunately Mr. Jan Duin refuses to have any contact with me, for reasons I can’t ex[plain. I’ve tried connecting him through his distant ancestor, but it was a patronymic and I could find no connection, even though I have compiled a database of 6,000 descendants from my Pieter. Most recently, I have located a female descendent of a Finnish Blankman family, but there are no male survivors alive to test. She and I are meeting in Finland this June, and although she now has her Family Finder test results, I am not one of her matches, but several of her matches are also my matches, mostly through the Y-Chromosome, which, of course, she could not test. It seems highly likely that Jan Duin’s line would tie into my Blankman line if he would only cooperate. It is just hard for me to understand why somebody would go to the expense to have tests performed, but then not be interested in the only match he probably has?

  3. Diane (Terpstra) Quale says:

    I had my DNA tested through Ancestry. My father’s parents immigrated from Friesland and my mother’s from Germany. The only surprise was 5% from Ireland. I had my Aunt (only sibling left of my father) tested and she is 49% Scandinavia, 46% Europe West, and 2% Ireland. Is so much of the Scandinavian odd in relationship to Friesland?

    • I would not be surprised if Frisian DNA is quite different to the rest of the Netherlands, since they are culturally a different people with their own language. So unless there were many Frisians in the reference sample, Ancestry would not be able to ID them properly. Contacts across the North Sea might explain why they resemble Scandinavian DNA more than Western European DNA. Friesland was also a frequent target for Vikings.

      • @Diane try downloading your information to the free website – gedmatch. They seem to have Frisian DNA. I say seem because my family is mostly western southern German, Swiss, and French and I get an outrageously high amount of Frisian DNA which might be due more to the French / Huguenot connection than actual authentic Frisian for myself. Give it a shot. If you find that you appear to have little Frisian DNA on that website than their reference samples might just be as I have assumed and as “hot air” worthy as 23&me’s idea to genetic ethnicity.

        But a word of advice unless the ethnicity is giving you something completely alien take it with a large grain of salt. Most of these ethnicity websites can’t differ between major populations [French, German, etc.] due to population movement. Probably doesn’t help I have heard their sample populations in almost every ethnic DNA estimate only reference up to their grandparents which is pretty meaningless…. I mean someone with an English sounding name puts “English” down when two generations further back their ancestors were Russians for example.

        I have yet to find a ethnic DNA website that is accurate to my known heritage. Close sure but they keep referring to me as “Hungarian” or “Croatian” or “Russian” or even “Serb” when the furthest east ancestors ever went was Berlin. But they can’t for the life of them identify the Japanese great-grandfather nor the Ashkenazi beyond minimal.

      • I got a small percentage of Scandinavian that surprised me in my AncestryDNA results. My maternal grandmother just took the test and got 35% Scandinavian, though she has done genealogy all the way back to the late 1500s of ancestors who all lived in Friesland. I had been wondering if the Frieslanders had different ethnic influences in their DNA than the Dutch do and they may test as Scandinavian? If anybody has any information that would be helpful, my grandmother and I would appreciate. 🙂

        • Robert Andersson says:

          I dont know if I can give you any help but
          I hail from south of Sweden and I can trace my roots which are Swedish and Danish at least 500 years back in time.
          So I expecetd when I had my DNA tested to find 100% Scandinavian DNA. To my surprise I have 30% Dutch DNA probably from Friesland. I have later found that a lot of South Scandinavians have Dutch DNA. It hasnt got to do with anything from viking age but later around 1200-1400.At that time many Dutch traders and people within shipping moved to places like Gothenburg, Malmo Copenhagen. At the same time Scandinavians moved to Friesland.

  4. Regarding 23andMe my experience is similar to yours.

    According to their present Ancestry Composition analysis I would be even 31% British Irish, while paper records (regarding all ancestors) show no such ancestry down to at least the 17th century.
    Apparently, 23andme messed things up by including customer data, which are only 2 generations deep. I.e.: I seem to relate to many 23andme participants who report to have grandparents from Britain or Ireland, but I actually relate to the ancestors of these grandparents, in mainland Europe (who go unnoticed by 23andme).

    Due to US regulations the 23andme earning model collapsed a few years ago and in their effort to repair the damage they now seem to concentrate on health related data from North American participants ONLY (data they can sell!). Although they have started web offerings outside de US, the experience with the US main office so far does not bode well. Very unfortunate, as the huge 23andme database is certainly also of interest from the European perspective (and from the international perspective in general).

    Would you know of any serious European dna ancestry research initiative?

  5. Leg eens uit wat je bedoeld met Brits DNA en continentaal West Europees DNA? 99.9% van het dna van bevolkingen uit West Europa komen overeen. Er is heel veel overlap. Mensen met een Duitse achtergronden scoren heel hoog voor Groot Brittanie op ancestry en andersom zijn er Britten die heel laag scoren voor Groot Brittanie en met ‘Europe West’ eindigen. Het dna lijkt te veel op elkaar om van elkaar te kunnen worden onderscheiden.

  6. https://support.ancestry.com/s/question/0D51500001rVa9PCAS/dna-results-showing-great-britain-when-no-ancestors-are-known-to-be-from-there?t=

    Interessante discussie gaande op het ancestry forum aangaande het feit dat de verschillende populaties in West Europa genetisch gezien moeilijk van elkaar zijn te onderscheiden, dit verklaart voor een deel waarom er zoveel mensen met Duitse of Scandinavische voorouders tot hun verbazing “Britse resultaten” hebben. Brits DNA is, op Wales na, heel divers. In Noord Engeland en Schotland zijn er veel Scandinavische invloeden te bespeuren en in het oosten van Engeland wijkt het DNA nauwelijks af van dat van Friesland, wat niet zo verwonderlijk is, aangezien het laagland van Engeland, de zogenaamde ‘Fens’ in Norfolk en Suffolk, een voortzetting is van de Noord Europese Laagvlakte en gedurende de laatste Ijstijd gewoon aan Nederland vastzat.

    Het feit dat iemand een hoog percentage Brits of Scandinavisch DNA heeft als men zich autosomaal laat testen bij ancestry of 23andme, wil nog niet gelijk zeggen dat de voorouders daadwerkelijk uit deze gebieden kwamen. Het zeg meer iets over het feit dat het genetisch gezien heel moeilijk is deze populaties van elkaar te onderscheiden omdat er zowel geografisch als historisch gezien een grote overlap bestaat tussen deze regios. Jouw DNA vertoont om deze reden grote overeenkomsten met de Britse proefpersonen uit het Ancestry DNA database, net zoals bij veel anderen hun DNA juist overeenkomsten vertonen met de Scandinavische proefpersonen uit het Ancestry database.

  7. Ancestry biedt de mogelijkheid aan het ruwe DNA materiaal te uploaden naar Gedmatch. Gedmatch maakt gebruik van andere algoritmes dan ancestry.com en 23andme. Daarnaast biedt Gedmatch de mogelijkheid het DNA te vergelijken met verschillende populaties, zowel op Europees niveau als op wereld niveau. Op deze manier is het mogelijk te achterhalen bij welke populaties een individu clustert.

    Gedmatch maakt op mij een meer wetenschappelijke indruk en werkt daarnaast op non-profit basis. Het probleem met bedrijven die louter op commerciële basis functioneren, zoals ancestry, is dat het wetenschappelijke aspect wordt ondergesneeuwd. Ancestry is voornamelijk gericht op een Amerikaans publiek die willen weten waar hun voorouders vandaan kwamen, aangezien Amerika van oudsher een ‘melting pot’ is van verschillende etnische groepen. Commerciële bedrijven als Ancestry springen hierop in. Het feit is echter dat verschillende populaties genetisch gezien niet van elkaar te onderscheidden zijn, generaliserende categorieën als ‘Great Britain’ of ‘Europe West’ zijn daarom zinloos. Ancestry geeft bijvoorbeeld ook aan dat categorieën als ‘Great Britain’ aangetroffen worden in andere delen van Europa. Mijns inziens is Gedmatch superieur aan ancestry. Daarnaast kan een mitochondriaal (X-chromosomen test) of een Y-chromosomen test uitsluitsel bieden.

  8. Dutch DNA is in itself likely to be composite of Germanic, Anglo-Saxon, Scandinavian, and so on. The fact 23andme and FTDNA list the various subtypes as subtypes rather than the composite – i.e. Dutch – has to do with the levels of analysis the company chooses to use. They are not ‘missing the mark’ – Dutch DNA is simply a genetic composite, and therefore breaking it down into its constituent components makes sense.

    Second, we tend to link DNA categorisations to national borders (and we shouldn’t). You state that 20% British-Irish DNA is way too high, and that’s because you’re looking at the figure as determining where ‘one is from’. DNA does not respect the artificial constructs that are modern-day national borders. A significant proportion of modern-day Brits and Irishmen are descended from Anglo-Saxon migrants from continental Europe who moved there over a thousand years ago. There will therefore be close genetic links between populations on either side of the Channel. The 28-something percent, then, is genetically conserved material tracing back to these common ancestors. It does not make you British-Irish; rather, it means that Dutch DNA is to a large extent Anglo-Saxon.

Leave comment

*