If your reason for wanting to take a DNA test is to learn about your ancestral heritage, then save your money, because they are mostly useless in that regard. DNA ancestral heritage tests, often called admixture reports or ethnicity estimates, are the “wild west” of the DNA world. We are given to believe that DNA science is an extremely exacting science–and it is. However, when it comes to revealing your ancestral heritage, you might as well be getting your information from a snake oil salesman.
This article is Part 2 of my DNA Testing Case Study. It compares 10 different ancestral heritage admixture reports, including those from all the major DNA testing companies. This is an actual case study, conducted on the same identical DNA (my DNA). However, it resulted in extremely wide variations of reported ancestral heritage, heritage that even changed dramatically over time.
In Part 1, DNA Testing: A Case Study With Staggering Conclusions, I compared the DNA matching results from six different companies. DNA testing is excellent at that particular task — identifying DNA relatives, and that article reveals a clear winner in the DNA matching sweepstakes. It also provides great recommendations and suggestions on how you can save money by taking just one test and then uploading your DNA to other testing and analysis sites for free.
DNA ancestral heritage is the most widely marketed feature of many DNA testing kits, and this article clearly reveals why it might also be the most useless feature of DNA testing. The reports they are selling you are so sloppy and inaccurate that the Federal Trade Commission should force them to disclose “For Entertainment Purposes Only” in bold lettering.
A Little Background
Heritage and ethnicity are the words most often used by the general public, but “admixture” is the technical term that describes our genetic heritage. Genetic mixtures occur when two distinct groups combine, and their offspring is now a blend (mixture) of the two heritages. For example, if a person of Scandinavian heritage had a child with a person of German heritage, the child would have an admixture of Scandinavian and German.
The DNA testing company’s marketing propaganda wants you to believe their DNA test will tell you if you are a Viking or Italian. However, it will not. Instead, it is going to report your heritage mixture, based on a combination of all your ancestors and their proprietary analysis of your DNA. You have four grandparents, eight great-grandparents, sixteen 2x great, thirty-two 3x great, and sixty-four 4x great-grandparents. You are a mixture of all 64 of your 4x great-grandparents (and each of the generations that came before them), and on average, each of these 64 ancestors contributes about 1.6% of your DNA.
If one or two of these 64 ancestors did indeed descend from Vikings, then your heritage report might say 1%-3% Scandinavian. Now, what are you going to do with that information besides “brag” about it? Which of the 64 were the contributors? How is this going to help you build your tree? What about your 23% German heritage? But this case study raises an even more important question — what if the report is flat-out wrong?
Each DNA company uses different reference points for their “pure” heritage samples, and they are constantly revising their analysis algorithms. As a result, you may be 20% Irish today and 34% Irish next week or whenever they next update their programs. The comparison table below has some great examples of this. Three different versions of Ancestry’s reports and two versions of FamilyTreeDNA’s reports are shown. Both companies recently updated their algorithms, and the current results are extremely different from what they were a couple of months ago (using the same DNA sample).
Report Comparisons & Summary
Here is a comparison of my DNA ancestral heritage admixture reports from the major testing companies. Please note that all the differences between each report are the result of analysis differences at each firm. My DNA was the one constant in this experiment/process. And since this is “For Entertainment Purposes Only,” I went ahead and added one more report for your entertainment — the last column on the right is my heritage self-analysis, based on my family tree (no DNA).
Looking at the first line in this comparison table, you can see that somewhere between 5.0% and 89.9% of my heritage is from the British Isles (UK and Ireland). Thank you DNA testing companies, that really pinpoints it for me. Additionally, my heritage is apparently much different today than it was only a couple of months ago. FamilyTreeDNA said my UK & Ireland heritage was just 5% in August, but it jumped to 48% in September. It must have been that English gin I began drinking in September.
Speaking of changes, look at the sub-categories for the UK. According to Ancestry, I went from being 9% Scottish to 52% Scottish overnight. I didn’t feel a thing, and haggis still fails to make my list of top-10 favorite foods.
My Scandinavian heritage lies somewhere between 0% (MyHeritage and LivingDNA) and a whopping 23% (FamilyTreeDNA). I’m sure glad these high-priced companies were able to nail that one down. My self-analysis puts it at 0% because I have not found any Scandinavians in my family tree. However, current genetic theory suggests that Scotland, England, and Ireland all had Norwegian settlers (invaders) thousands of years ago, so if you look back in time far enough, some Scandinavian heritage for me makes sense.
Southern Europe is another one of my heritage sources that all companies can agree on — NOT. The GEDmatch Eurogenes K13 Model puts it at 24.7%, while the majority of other reports place it at zero. The GEDmatch report also introduced another category not used by the other companies — Baltic. And with a huge 21% Baltic allocation, they apparently believe it is not a fluke. While we are on the subject of the GEDmatch analysis, they also claim part of my heritage is from three other locations that do not show up in other reports. They put me at 3.5% Asian, 1.1% Australian Aborigine, and 0.1% Siberian. My guess is that the anomalies rising from the GEDmatch report have to do with them using ancient reference samples instead of ones that are merely hundreds of years old.
Last, but not least, are my African heritage variations. The majority of the reports put it at zero, three have trace results, while MyHeritage proclaims it to be quite large at 6.5%, I should point out here that MyHeritage breaks it down further to 5.4% North Africa and 1.1% Nigeria. Once again, I believe some of this can be attributed to how far back the reports are looking. After all, everyone on the planet was thought to be 100% African about 75,000 years ago.
Perhaps one day these testing companies will get closer to getting it right, or at least closer to agreeing with each other. However, at this time, DNA ancestral heritage testing and reporting is a complete farce.
Individual Reports
Here are the detailed DNA ancestral heritage reports used in the comparison table above. The August 2020 versions of these reports are displayed, along with comments about subsequent updates.
Ancestry DNA Story: Ethnicity Estimate of August 2020
Ancestry’s ethnicity estimate report indicates 67% of my heritage is from England, Wales, & Northwestern Europe. Another 18% comes from Ireland & Scotland. I believe my Rowland surname ancestors arrived in Pennsylvania in the early 1700s from somewhere in the British Isles. However, this report does not help me narrow it down. The mix also shows I have 7% Germanic, 4% Norway, 2% France, 1% Italy, and 1% Eastern Europe heritage.
Ancestry has another feature called “Additional Communities,” and I have been identified has having many European ancestors that originally settled in New York. No other “communities” for me have been identified.
The scientists and programmers at Ancestry have updated their process four or five times since I first tested with them in 2016. This chart shows my August 2020 results, and the summary table in the section above also includes my first 2016 report along with my newest September 2020 report. Apparently, a large percentage of test-takers saw their Scottish heritage explode with this latest update from Ancestry. As such, I fully expect the next update from Ancestry will show a large reduction of Scottish heritage for me and many others.
FamilyTreeDNA Ancestral Origins of August 2020
The FamilyTreeDNA Ancestral Origins report states I am 71% West & Central Europe, 23% Scandinavian, and less than 1% North & Central America (Native American). According to them, I am much more Viking, a lot less British, and a little bit Native American versus what Ancestry is telling me.
Like Ancestry, FamilyTreeDNA also revised its algorithms recently. Their new “Origins Version 3” results are vastly different from their August report. My Western Europe heritage plunged from 71% to just 40%, with nearly all of that increasing my UK & Ireland heritage from 5% to 48%. From what I believe to be true about my heritage, this revision seems to have moved in the correct direction, but a nearly 10x jump from 5% to 48% is unsettling.
FamilyTreeDNA also provided me with an “Ancient European Origins” report. This report states that my ancient heritage is:
- 45% Farmer
- 43% Hunter-Gatherer
- 12% Metal Age Invader
MyHeritage Ethnicity Estimate
The MyHeritage Ethnicity Estimate puts my heritage at 79% England, 10.9% Ireland-Scotland-Wales, 3.6% Iberian peninsula, 5.4% North Africa, and 1.1% Nigerian. The African heritage figure of 6.5% is far greater than other reports which typically range from 0% to 1%. Additionally, my French and German heritage has disappeared.
I’m at a loss to explain the large African estimate. Their map for Iberian heritage includes France and Italy. However, I am surprised at the complete lack of any German heritage being reported. As they continue to build their database and improve their algorithms, I expect they will be providing revised reports in the future.
23andMe Ancestry Composition Report
The 23andMe Ancestry Composition report increases my French and German heritage to nearly 22% versus the 0% in the MyHeritage report. It also gives me back my British heritage that the FamilyTreeDNA report took away.
They assigned a large 22.6% to a “Broadly Northwestern Europe” category that they apparently were not able to narrow down and include in either the French & German or the British & Irish categories.
Digging deeper into the reported “Trace” ancestry, the report reveals I have 0.4% Nigeria and 0.3% Native American heritage.
23andMe claims to include 224 population samples in their analysis. My DNA is compared to each and then categorized as being detected or not detected. If detected, then it is often further categorized and labeled as being a possible match, a likely match, or a highly likely match.
This report was generated with their Ancestry Composition v5.2 software and genotyping chip version 5 (last updated October 29, 2019).
Although Y-DNA and mtDNA reports often require a unique test and/or payment, 23andMe includes haplogroup reports. My paternal haplogroup is stated to be R-P311 and my maternal haplogroup is H3. This will be discussed further in a future Y-DNA article, but the R-P311 group traces back about 10,000 years, or 400 generations ago, and about 5.5% of 23andMe customers are in this group. My maternal haplogroup of H3 goes back about 9,000 years, and about 2% of all 23andMe customers are in this group.
23andMe Neanderthal Report
As a bonus, 23andMe also provided a report on my Neanderthal DNA. In August, I had 304 Neanderthal variants, which is apparently well above average among 23andMe customers. I’m not sure if I am supposed to be thrilled or concerned about these facts. My reported 4% Neanderthal DNA is at about the same level as my Viking DNA (Norway). I believe this is my “free pass” to dress as a Neanderthal Viking for Halloween and not be accused of cultural appropriation.
Like other heritage reports, my Neanderthal heritage appears to be changing over time. My report has been updated to show my quantity of Neanderthal variants dropping from 304 to 254, and now amounts to only about 2% of my DNA. Apparently, this is still above average, as it is a higher figure than 72% of their other customers.
By the way, the Neanderthal species became extinct about 40,000 years ago. So yes, DNA can see very deep into the past.
LivingDNA Recent Ancestry
The LivingDNA Recent Ancestry report goes much further than the others in segmenting my DNA ancestral heritage by regions of the UK and Ireland. This is not surprising since they are a UK-based company. The largest chunk in this subgroup is the 15.7% designated as Northern Ireland and Southwest Scotland. If you are familiar with the history of this region, then you know this is a result of the Ulster Plantations, or England’s attempt to get rid of the unruly Scots by shipping them to Ulster province in Northern Ireland. For US-based genealogy, this group of people are typically referred to as Scots-Irish. In the UK, they are more commonly called Ulster-Scots. In the summary table near the top of this article, I put half into the 15.7% combination into the Scotland category and the other half into Northern Ireland.
My second largest UK segment is Lincolnshire, followed by Central England, Cumbria, and Southeast England. Northumbria, Cornwall, South Wales, and Devon are among the many other regions included in my mixture. It would appear that I am quite the melting pot of UK heritage.
It is interesting to note that my South Germanic heritage jumps to more than 26% in this report, which fits with my understanding that one of my four grandparents was full German. Not to be overlooked, this report also puts me at 1% Arabian, and I cannot explain that.
Like 23andMe, LivingDNA also provides Y-DNA and mtDNA reports. LivingDNA reports my paternal haplogroup as R-P312, which is a subclad of the R-P311 group reported by 23andMe. This means the LivingDNA report is a little more precise. I have had extensive Y-DNA testing performed, and I know that my actual haplogroup is R-FT65023, which is a subset of R-P312.
LivingDNA reports my maternal haplogroup as H3, which matches that of the 23andMe report.
GEDmatch Admixture
At GEDmatch, there are dozens of possible admixture reports you can run, and they are all free. Deciding which one to run can be overwhelming. Perhaps someone has written a guide, but I just go in and try the different tests and models. One will even tell you the amount of your pygmy heritage, and it reports mine at 0%, which seems to agree with my current 6’4″ height.
The report shown to the left is from the Eurogenes Project K13 Model, which uses the 13 shown genetic population definitions.
Of particular note is that this is the first appearance of any Asian heritage for me, with my Asian DNA reported to be 3.5%. This is further broken down into West Asian at 2.7% and East Asian at 0.8%. The 1.1% Oceanian Aboriginal and 0.1% Siberian are also new and come as big surprises.
It is also not obvious to me where this model draws its geographic lines (and their link to its descriptions is broken), so I do not know exactly how it defines Baltic. I tend to think of the Baltic region as being Eastern Europe, but perhaps their definition stretches more westward into modern-day Germany.
The author of this particular model focuses much of his research on “ancient” DNA, and therefore these comparisons are probably looking thousands of years deeper into my DNA than the more recent ancestral reports of the other companies.
My Self-Generated Ancestral Heritage Report
I have a fairly extensive and well-researched family tree. This makes it possible for me to perform my own analysis and generate my own report.
My maternal grandmother was the easiest to analyze, as all of her ancestors can be traced to Germany. Therefore, I would expect her 100% German heritage to translate to about 25% German for me. It is important to note here that DNA transmission is somewhat random, so this “about 25%” can range from 17% to 32%. This is covered more extensively in Genetic Genealogy.
My paternal grandmother was born in Canada, and two of her grandparents can be traced back to Devon, England. Her other two grandparents are a mixture of French, English, and Irish, although many generations were in North America. If I assign equal weight to these three, then this will boost her English heritage to 66.6%, and put her French and Irish portions at about 16.6% each.
My maternal grandfather is probably a textbook blend of Scots-Irish, Scotland, and England, although I believe all 16 of my 4x great-grandparents on his side were living in Virginia in 1800. Assigning equal weight to these three groups, my maternal grandfather’s heritage is estimated at 33.3 % each of Scots-Irish (North Ireland), Scottish, and English.
My paternal grandfather’s heritage has an interesting history that includes a Dutch immigrant marrying a Native American on Long Island in the early 1600s. Although this is well-documented, it is 11 generations away from me, and my Native American DNA might not be enough to detect. Although my grandfather’s mother was born with the Dutch surname, the Dutch portion of his heritage was watered down with marriages to various German and English spouses through the generations. My estimate for my paternal grandfather is 50% German, and 12% each of Dutch, Scots-Irish, Scottish, and English. I applied the remaining 2% to my Native American heritage.
The next step was to add up all the various heritage estimates of my grandparents and divide them by four. My self-analysis suggests I am 99.5% European and 0.5% New World. My European component is further broken down to 54.9% British Isles and 44.7% for continental Europe. I knew I was 25% German from my maternal grandmother, but I was surprised the ancestors of my other grandparents boosted this number to 37.5%. My 0.5% estimate of my Native American heritage seems to agree with the DNA reports that place it between 0% and 1.2%.
Who is to say my self-generated report is any better or worse than the paid professional reports? So far, I have not followed in the footsteps of the DNA company reports by issuing updates that are vastly different from this one.
DNA Ancestral Heritage Report Case Study Conclusions and Recommendations
- DNA ancestral heritage admixture reports being sold by the major DNA testing companies appear to be a complete farce, at least when it comes to my DNA. They don’t come close to agreeing with reports from the other companies, and they don’t even agree with the prior reports issued by the same company.
- If your reason for wanting to purchase a DNA test is to learn about your heritage, then you will probably be wasting your money if you are primarily of European descent. If you are a multi-continent blend, then perhaps they might be more enlightening.
- If you are adopted or have little or no clue as to your biological heritage, then a heritage test may provide some insight. However, do not put too much faith into reports that can change a given heritage from 9% to 52% in a single update.
- You can save a lot of money by doing a DNA test with AncestryDNA, downloading your raw DNA data, and then uploading it free to FamilyTreeDNA, MyHeritage, LivingDNA, and GEDmatch as described in my DNA Testing: A Case Study With Staggering Conclusions article.
- These heritage reports should come with a warning that they are for “For Entertainment Purposes Only.” I will admit to being totally entertained by the ridiculous inconsistencies, variations, and changes of the reports I paid for. However, your mileage may vary.
Suggested Further Reading
Genetic Genealogy – an introduction to DNA, DNA testing, and the randomness of generational DNA transmission.
Genealogy Websites – a summary of the companies and websites discussed in this article.
Home Page – Rowland Genealogy is much more than just a tree, and it is for much more than just Rowland families. Take a look.
——
If you are a Rowland or have a Rowland in your family tree, then you need to subscribe to the Rowland Genealogy Newsletter.
Follow RowlandGenealogy on Facebook to keep abreast of new activities and posts.
My DNA tests are accurate. I tested with Living DNA, and then I uploaded my results to MyHeritage. Both companies said the same thing.
It is harder to decipher Northwest Europeans because they cluster so close to each other. English, Germans, Dutch and Swedes etc all overlap with one another, and that is why your results look so varied, from one company to the next. Maybe these companies should lump Northwest Europe into one category, and be done with it. I find Northwest Europeans complain about their results much more than other people.
Your matches are what really confirms your ethnicity. For example, you may be English, and score Scandinavian instead of English, but if you get lots of English matches, that proves you are English. You maybe German, and score English instead of German, but if you get lots of German matches, that proves you are German.
Nick, thanks for reading and sharing your contextual commentary on this subject.
Good article Ron. My experience is much the same but the varying Heritage results did make sense in my case. The closest to my Tree is Living DNA which has me down for recent heritage as British from Suffolk and long term as Scandinavian. The others range from 3% British to 50% with the balance as Scotland, Ireland, Scandinavia in varying amounts.
I am British and my family tree for 500 years is predominantly North Essex, Suffolk (the majority) and Norfolk. Luckily for me, my families hardly moved from their home towns/villages for hundreds of years which made my Genealogy research reasonably easy to pin down. It was supported by family folklore passed down by word of mouth.
As you probably know, these counties are in the English region called East Anglia (Land of the Eastern Angles) which is the big “bump” on the Eastern side of England. After the Roman empire collapsed around 450, this area saw an influx of people called “Angles” from the Southern part of the Jutland peninsular (just below modern day Denmark). The Angles homeland was South of the Jutes in Jutland and North of the Saxons. After they settled the Danes/Vikings arrived about 400 years later and following that the area was ruled by them (Danelaw) almost up to the Norman invasion.
The resulting DNA from these post Roman settlers would show very similar results to people in modern day Scandinavia, Scotland and Ireland, which is where the DNA companies struggle as you rightly pointed out, Ron. Their samples are all pretty much based on where people live today rather than where they originated.
This is why my DNA has such wild heritage variations. I think Living DNA must overlay a bit of historical knowledge to modify the result, they are a UK company so that would make sense. I imagine this is how they can split the results between recent and longer term for the Y and Mitochondrial and be more regionally accurate.
Anyway, I suspect your result might indicate a reasonable Celtic content because of the multi-county result. The Celts were the people in England before the Romans invaded, they didn’t go anywhere but became what is known as Romano-British, living in the style of Romans. After the Romans left they returned to a similar way of life to before the Roman invasion, they lived side by side and mixed with Angles, Saxons, Vikings etc but weren’t displaced very much. As a result their DNA will still exist quite strongly all around England and won’t be particularly regional.
I imagine that I descend largely from the Settlers and you are more of a mix of Original Celtic British and post Roman Settlers.
Any way that’s my take on it all…… for the moment 🙂
Cheers
Nick