Recently, Family Tree DNA introduced their new ethnicity tool, myOrigins as part of their autosomal Family Finder product. This means that all of the major players in this arena using chip based technology (except for the Genographic project) have now updated their tools. Both 23andMe and Ancestry introduced updated versions of their tools in the fall of 2013. In essence, this is the second generation of these biogeographical or ethnicity products. So lets take a look and see how the vendors are doing.
In a recent article, I discussed the process for determining ethnicity percentages using biogeographical ancestry, or BGA, tools. The process is pretty much the same, regardless of which vendor’s results you are looking at. The variant is, of course, the underlying population data base, it’s quality and quantity, and the way the vendors choose to construct and name their regions.
I’ve been comparing my own known and proven genealogy pedigree breakdown to the vendors results for some time now. Let’s see how the new versions stack up to a known pedigree.
The paper, Revealing American Indian and Minority Heritage Using Y-line, Mitochondrial, Autosomal and X Chromosomal Testing Data Combined with Pedigree Analysis was published in the Fall 2010 issue of JoGG, Vol. 6 issue 1.
The pedigree analysis portion of this document begins about page 8. My ancestral breakdown is as follows:
|European by DNA||6.8362|
This leaves about 25% unknown.
Let’s look at each vendor’s results one by one.
My results using the speculative comparison mode at 23andMe are shown in a chart, below.
|23andMe Category||23andMe Percentage|
|British and Irish||39.2|
|Nonspecific North European||27.9|
|Nonspecific South European||1.6|
|Nonspecific East Asian/Native American||0.1|
|Middle East/North Africa||0.1|
At 23andMe, if you have questions about what exact population makes up each category, just click on the arrow beside the category when you hover over it.
For example, I wasn’t sure exactly what comprises Eastern European, so I clicked.
The first thing I see is sample size and where the samples come from, public data bases or the 23andMe data base. Their samples, across all categories, are most prevalently from their own data base. A rough add shows about 14,000 samples in total.
Clicking on “show details” provides me with the following information about the specific locations of included populations.
Using this information, and reorganizing my results a bit, the chart below shows the comparison between my pedigree chart and the 23andMe results. In cases where the vendor’s categories spanned several of mine, I have added mine together to match the vendor category. A perfect example is shown in row 1, below, where I added France, Holland, Germany and Switzerland together to equal the 23andMe French and German category. Checking their reference populations shows that all 4 of these countries are included in their French and German group.
|Geography||Pedigree Percent||23andMe %|
|Germany, Holland, Switzerland & France||45.7451||15.6|
|Native American||0.2933||0.4 (Native/East Asian)|
|Turkish||0.0031||0.1 (Middle East/North Africa)|
|European by DNA||6.8362||4.9 (nonspecific European)|
|Unknown||25||27.9 (North European)|
I can also change to the Chromosome view to see the results mapped onto my chromosomes.
The 23andMe Reference Population
According to the 23andMe customer care pages, “Ancestry Composition uses 31 reference populations, based on public reference datasets as well as a significant number of 23andMe members with known ancestry. The public reference datasets we’ve drawn from include the Human Genome Diversity Project, HapMap, and the 1000 Genomes project. For these datasets as well as the data from 23andMe, we perform filtering to ensure accuracy.
Populations are selected for Ancestry Composition by studying the cluster plots of the reference individuals, choosing candidate populations that appear to cluster together, and then evaluating whether we can distinguish the groups in practice. The population labels refer to genetically similar groups, rather than nationalities.”
Additional detailed information about Ancestry Composition is available here.
Ancestry is a bit more difficult to categorize, because their map regions are vastly overlapping. For example, the west Europe category is shown above, and the Scandinavian is shown below.
Both categories cover the Netherlands, Germany and part of the UK.
My Ancestry percentages are:
|Ancestry Category||Ancestry Percentage|
Below, my pedigree percentages as compared to Ancestry’s categories, with category adjustments.
|Geography||Pedigree Percent||Ancestry %|
|West European||52.584 (combined from below)||79|
|European by DNA||6.8362||Combined|
|Native American||0.2933||~1 incl East Asian|
|Turkish||0.0031||1 (North Africa)|
Ancestry’s European populations and regions are so broadly overlapping that almost any interpretation is possible. For example, the Netherlands could be included in several categories – and based up on the history of the country, that’s probably legitimate.
At Ancestry, clicking on a region, then scrolling down will provide additional information about that region of the world, both their population and history.
The Ancestry Reference Population
Just below your ethnicity map is a section titled “Get the Most Out of Your Ethnicity Estimate.” It’s worth clicking, reading and watching the video. Ancestry states that they utilized about 3000 reference samples, pared from 4245 samples taken from people whose ethnicity seems to be entirely from that specific location in the world.
You can read more in their white paper about ethnicity prediction.
Family Tree DNA’s myOrigins
I wrote about the release of my Origins recently, so I won’t repeat the information about reference populations and such found in that article.
Family Tree DNA shows matches by region. Clicking on the major regions, European and Middle Eastern, shown above, display the clusters within regions. In addition, your Family Finder matches that match your ethnicity are shown in highest match order in the bottom left corner of your match page.
Clicking on a particular cluster, such as Trans-Ural Peneplain, highlights that cluster on the map and then shows a description in the lower left hand corner of the page.
Family Tree DNA shows my ethnicity results as follows.
|Family Tree DNA Category||Family Tree DNA Percentage|
|European Coastal Plain||68|
|European Coastal Islands||7|
|Anatolia and Caucus||3|
Below, my pedigree results reorganized a bit and compared to Family Tree DNA’s categories.
|Geography||Pedigree Percent||Family Tree DNA %|
|European Coastal Plain||45.7478||68|
|British Isles||22.6104||7 (Coastal Islands)|
|Turkish||0.0031||3 (Anatolia and Caucus)|
|European by DNA||6.8362|
Third Party Admixture Tools
www.GedMatch.com is kind enough to include 4 different admixture utilities, contributed by different developers, in their toolbox. Remember, GedMatch is a free, meaning a contribution site – so if you utilize and enjoy their tools – please contribute.
On their main page, after signing in and transferring your raw data files from either 23andMe, Family Tree DNA or Ancestry, you will see your list of options. Among them is “admixture.” Click there.
Of the 4 tools shown, MDLP is not recommended for populations outside of Europe, such as Asian, African or Native American, so I’ve skipped that one entirely.
I selected Admixture Proportions for the part of this exercise that includes the pie chart.
The next option is Eurogenes K13 Admixture Proportions. My results are shown below.
Of course, there is no guide in terms of label definition, so we’re guessing a bit.
|Geography||Pedigree Percent||Eurogenes K13%|
|British Isles||22.6104||Combined above|
|European by DNA||6.8362||Combined above|
|Native American||0.2933||2.74 combined East Asian, Siberian, Amerindian and South Asian|
|Turkish||0.0031||1.78 Red Sea|
Next is Dodecad K12b
According to John at GedMatch, there is a more current version of Dodecad, but the developer has opted not to contribute the current or future versions.
By the way, in case you’re wondering, Gedrosia is an area along the Indian Ocean – I had to look it up!
|Geography||Pedigree Percent||Dodecad K12b|
|British Isles||22.6104||Combined above|
|European by DNA||6.8362||Combined above|
|Native American||0.2933||3.02 Siberian, South Asia, SW Asia, East Asia|
Third is Harappaworld.
Baloch is an area in the Iranian plateau.
|Geography||Pedigree Percent||Harappaworld %|
|British Isles||22.6104||Combined above|
|European by DNA||6.8362||Combined above|
|Native American||0.2933||2.81 SE Asia, Siberia, NE Asian, American, Beringian|
The wide variety found in these results makes me curious about how my European results would be categorized using the MDLP tool, understanding that it will not pick up Native, Asian or African.
The Celto-Germanic category is very close to my mainland European total – but of course, many Germanic people settled in the British Isles.
Second Generation Report Card
Many of these tools picked up my Native American heritage, along with the African. Yes, these are very small amounts, but I do have several proven lines. By proven, I mean both by paper trail (Acadian church and other records) and genetics, meaning Yline and mtDNA. There is no arguing with that combination. I also have other Native lines that are less well proven. So I’m very glad to see the improvements in that area.
Recent developments in historical research and my mitochondrial DNA matches show that my most distant maternal ancestral line in Germany have some type of a Scandinavian connection. How did this happen, and when? I just don’t know yet – but looking at the map below, which are my mtDNA full sequence matches, the pattern is clear.
Could the gene flow have potentially gone the other direction – from Germany to Scandinavia? Yes, it’s possible. But my relatively consistent Scandinavian ethnicity at around 10% seems unlikely if that were the case.
Actually, there is a second possibility for additional Scandinavian heritage and that’s my heavy Frisian heritage. In fact, most of my Dutch ancestors in Frisia were either on or very near the coast on the northernmost part of Holland and many were merchants.
I also have additional autosomal matches with people from Scandinavia – not huge matches – but matches just the same – all unexplained. The most notable of which, and the first I might add, is with my friend, Marja.
It’s extremely difficult to determine how distant the ancestry is that these tests are picking up. It could be anyplace from a generation ago to hundreds of generations ago. It all depends on how the DNA was passed, how isolated the population was, who tested today and which data bases are being utilized for comparison purposes along with their size and accuracy. In most cases, even though the vendors are being quite transparent, we still don’t know exactly who the population is that we match, or how representative it is of the entire population of that region. In some cases, when contributed data is being used, like testers at 23andMe, we don’t know if they understood or answered the questions about their ancestry correctly – and 23andMe is basing ethnicity results on their cumulative answers. In other words, we can’t see beneath the blanket – and even if we could – I don’t know that we’d understand how to interpret the components.
So Where Am I With This?
I knew already, through confirmed paper sources that most of my ancestry is in the European heartland – Germany, Holland, France as well as in the British Isles. Most of the companies and tools confirm this one way or another. That’s not a surprise. My 35 years of genealogical research has given me an extremely strong pedigree baseline that is invaluable for comparing vendor ethnicity results.
The Scandinavian results were somewhat of a surprise – especially at the level in which they are found. If this is accurate, and I tend to believe it is present at some level, then it must be a combined effect of many ancestors, because I have no missing or unknown ancestors in the first 5 generations and only 11 of 64 missing or without a surname in generation 6. Those missing ancestors in generation 6 only contribute about 1.5% of my DNA each, assuming they contribute an average of 50% of their DNA to offspring in each subsequent generation.
Clearly, to reach 10%, nearly all of my missing ancestors, in the US and Germany, England and the Netherlands would have to be 100% Scandinavian – or, alternately, I have quite a bit scattered around in many ancestors, which is a more likely scenario. Still, I’m having a difficult time with that 10% number in any scenario, but I will accept that there is some Scandinavian heritage one way or another. Finding it, however, genealogically is quite another matter.
However, I’m at a total loss as to the genesis of the South European and Mediterranean. This must be quite ancient. There are only two known possible ancestors from these regions and they are many generations back in time – and both are only inferred with clearly enough room to be disproven. One is a possible Jewish family who went to France from Spain in 1492 and the other is possibly a Roman soldier whose descendants are found within a few miles of a Roman fort site today in Lancashire. Neither of these ancestors could have contributed enough DNA to influence the outcome to the levels shown, so the South European/Mediterranean is either incorrect, or very deep ancestry.
The Eastern European makes more sense, given my amount of German heritage. The Germans are well known to be admixed with the Magyars and Huns, so while I can’t track it or prove it, it also doesn’t surprise me one bit given the history of the people and regions where my ancestors are found.
What’s the Net-Net of This?
This is interesting, very interesting. There are tips and clues buried here, especially when all of the various tools, including autosomal matching, Y and mtDNA, are utilized together for a larger picture. Alone, none of these tools are as powerful as they are combined.
I look forward to the day when the reference populations are in the tens of thousands, not hundreds. All of the tools will be far more accurate as the data base is built, refined and utilized.
Until then, I’ll continue to follow each release and watch for more tips and clues – and will compare the various tools. For example, I’m very pleased to see Family Tree DNA’s new ethnicity matching tool incorporated into myOrigins.
I’ve taken the basic approach that my proven pedigree chart is the most accurate, by far, followed by the general consensus of the combined results of all of the vendors. It’s particularly relevant when vendors who don’t use the same reference populations arrive at the same or similar results. For example, 23andMe uses primarily their own clients and Nat Geo of course, although I did not include them above because they haven’t released a new tool recently, uses their own population sample results.
National Geographic’s Geno2
Nat Geo took a bit of a different approach and it’s more difficult to compare to the others. They showed my ethnicity as 43% North European, 36% Mediterranean and 18% Southwest Asian.
While this initially looks very skewed, they then compared me to my two closest populations, genetically, which were the British and the Germans, which is absolutely correct, according to my pedigree chart. Both of these populations are within a few percent of my exact same ethnicity profile, shown below.
The description makes a lot of sense too. “The dominant 49% European component likely reflects the earliest settlers in Europe, hunter-gatherers who arrived there more than 35,000 years ago. The 44% Mediterranean and the 17% Southwest Asian percentages arrived later, with the spread of agriculture from the Fertile Crescent in the middle East, over the past 10,000 years. As these early farmers moved into Europe, they spread their genetic patterns as well.”
So while individually, and compared to my pedigree chart, these results appear questionable, especially the Mediterranean and Southwest Asian portions, in the context of the populations I know I descend from and most resemble, the results make perfect sense when compared to my closest matching populations. Those populations themselves include a significant amount of both Mediterranean and Southwest Asian. Looking at this, I feel a lot better about the accuracy of my results. Sometimes, perspective makes a world of difference.
It’s A Wrap
Just because we can’t exactly map the ethnicity results to our pedigree charts today doesn’t mean the results are entirely incorrect. It doesn’t mean they are entirely correct, either. The results may, in some cases, be showing where population groups descend from, not where our specific ancestors are found more recently. The more ancestors we have from a particular region, the more that region’s profile will show up in our own personal results. This explains why Mediterranean shows up, for example, from long ago but our one Native ancestor from 7 or 8 generations ago doesn’t. In my case, it would be because I have many British/German/Dutch lines that combine to show the ancient Mediterranean ancestry of these groups – where I have many fewer Native ancestors.
Vendors may be picking up deep ancestry that we can’t possible know about today – population migration. It’s not like our ancestors left a guidebook of their travels for us – at least – not outside of our DNA – and we, as a community, are still learning exactly how to read that! We are, after all, participants on the pioneering, leading edge of science.
Having said that, I’ll personally feel a lot better about these kinds of results when the underlying technology, data bases and different vendors’ tools mature to the point where there the differences between their results are minor.
For today, these are extremely interesting tools, just don’t try to overanalyze the results, especially if you’re looking for minority admixture. And if you don’t like your results, try a different vendor or tool, you’ll get an entirely new set to ponder!
I receive a small contribution when you click on some of the links to vendors in my articles. This does NOT increase the price you pay but helps me to keep the lights on and this informational blog free for everyone. Please click on the links in the articles or to the vendors below if you are purchasing products or DNA testing.
Thank you so much.
DNA Purchases and Free Transfers
- Family Tree DNA
- MyHeritage DNA only
- MyHeritage DNA plus Health
- MyHeritage FREE DNA file upload
- 23andMe Ancestry
- 23andMe Ancestry Plus Health
- Legacy Tree Genealogists for genealogy research
Pingback: Ethnicity Percentages – Second Generation Report Card | bittle30
Like you, I’m also concerned about the lack of consistency between the results of the various admixture tools, and the lack of a clear methodology to with which to evaluate them. In those cases where reference sample counts have been revealed, I can’t believe that the small number of samples could in any useful way “represent” the diversity and gene frequencies of a place like, say, Italy.
However, there’s another source of concern about the whole idea of quantifying ethnic origins. There has been some discussion in the blogs about the fact that the way recombination works, with at most a small number of crossovers on each chromosome per generation, will likely result in the loss of ALL genetic contribution from a significant number of ancestors, especially those farther back in the tree. Conversely, that fact also means that other ancestors will be over-represented in our DNA. So, when the pedigree says that I am 6.25% Eastern European Jewish, it is entirely possible that in fact I inherited no DNA whatever from my Prussian Jewish immigrant ancestor, or, on the other hand, that I happened to end up with a much larger amount. I haven’t worked out the confidence limits, but it seems to me that any of the “minor” components in our pedigrees must be subject to a huge amount of statistical error in terms of the contributions still found in our DNA from our ancestral populations. The “blockiness” of our DNA must be a strong reason why even a perfect admixture tool would not be expected to produce results that match our pedigrees very well. And that conclusion, in turn, makes it even more difficult to imagine how we can evaluate and verify the methodology of the admixture tools.
Pingback: Your Family Tree » Blog Archive » Ethnicity Explained
Took me a while to realize that you weren’t using my 23 and Me Ancestry Composition. Sure looks similar. My Asian/AmerIndian is from Yakut though.
These estimations have their problems, sometimes rooted in the migration of peoples prior to the middle ages and also earlier. For instance, in my compositions Ancestry says I am 45% British and 23 and Me lists me as 22.5% British and Irish. I have absolutely no British or Irish people in my pedigree and it goes back for at least 2 generations (sometimes 5-7, sometimes more) in their continental European homes. I posted a question about it with Ancestry at:
As far as I can tell the problem is due to the migration to England first of Belgae tribes prior to Roman occupation and second the later migrations of Saxons to England. Based in the Ancestry and the 23 and Me ancestry estimates I evidently (at least in my mind) carry a strong DNA signature of those tribes which did migrate to England, but from those Belgae and Saxons who remained in their homelands. My pedigree ancestry in Belgium is over 25% (although that overage is based in more Frankish Luxembourg province) and I also have a 25% Saxon ancestry from northwestern Germany.
My study of the numbers of people in Ancestry’s representative sample showed a decided bias against Western European DNA which would skew reported ancestry compositions away from those areas for those with populations which emigrated into England. Much in the same way that Ancestry had the “Scandavian Problem” because of a sampling bias, I now think they have a “British Problem.” At least for some of us.
One thing that might come from this if some DNA researcher is intersted is that using the DNA of Belgians and northwestern Germans we may be able to identify with some fair reliability the autosomal DNA signatures of Belgian Celtic and Saxon tribes. Who knows, maybe the 4th or 5th version of Ancestry DNA could result in generally identifying such general tribal ancestries.
Brava, Roberta! Another superlative analysis. I’ve also tested with all the companies, plus Geno2.0 and find the results so divergent that I have come to the conclusion that only broad generalizations can be accepted. I’m European; that’s about all I can conclude.
One thing I’d like to mention is the attenuated relationship between MyOrigins and FTDNA’s Ancestral Origins under their Y-DNA section. Given the very limited but now heavily studied scope of Y DNA, many people will look to see how this relates to the results from these autosomal tests. You’ve covered the subject in great detail in the past, but to give you an idea, Ancestral Origins shows Spain as the largest percentage, 1.7% of my matches at 12 STRs, 1 genetic distance, while MyOrigins shows absolutely zero contribution from Southern Europe which includes Iberia. Ancestry had 2.2% and 23andMe had a much larger 13%. Do I actually have Spanish ancestry? Who knows?
As you know, ancestral origins is the reported ethnicity of testers. Of course, the 12 markers results would go back further in time than the higher panels because they are less specific. 1.7% isn’t much, but it does make one wonder if that is a function of migration of those people to Spain, or something else. It’s all interesting to think about and to take into consideration as we look at other results and tools.
Exactly. As the King would say, “It is a puzzlement!”
Thank you so much, Roberta! This is the information that my cousins and I have been looking for and it has given me, with my limited knowledge, a much greater understanding. I took Native American artifacts to a local elementary school yesterday and we recorded questions and answers at the third grade level. We did not get into anything complicated, but working out the answers to the questions beforehand certainly had me checking and re-checking my sources and data. The media specialist has borrowed my materials including a copy of the 1733 Edward Moseley map. I feel very certain that I have a small percentage of Native American from the careful traditional genealogy done by one of my very astute and experienced cousins. For the many reasons mentioned above, I think it is unrealistic to expect find an exact percentage going back for so many generations. It appears that within two or three generations, the descendants of the Native American mother and English father had begun to intermarry and produce offspring. All I ever hope to demonstrate is that there WAS at least one Native American in my ancestry which would serve as supporting evidence for the traditional genealogy that has already been done. One of our cousins recently tested with 23andMe and reported that he is “so white he could pass for a sheet”. Some of our traditional genealogy would suggest that there was at least one Native American in his line of the family. We came from an area where our Native American ancestors were here to greet the first English settlers and the mixing would have occurred prior to the American Revolution by which time the majority of the Native Americans in this area had intermarried or sold their land and left the area to join with other groups of Native Americans. Our Yawpim reservation here was one of the earliest to be granted in the United States. Now our group of cousins have more information to ponder. As always, Roberta deserves thanks again.
Because Hatteras Island had a limited number of families and we know who some of them were who have a history of being Native, I would strongly encourage everyone there with an island heritage to test their Yline and mtdna so that we can eventually document who these ancestors were. The Basnett, Whidbey and Scarborough families from are surrounding Indian Town are the primary families that look to have been Native or “mixed” in the records. We have the Hatteras Y and mtDNA projects to track these families. Hatteras is a case where the Y and mtDNA may be more important than autosomal because we can build an island pedigree chart of the ancestral DNA of the earliest families.
I am new to this site and noticed your discussion of the Hatteras Native Americans. I have recently been researching my husbands Powers line which includes family surnames of Clark/e and Hardy, and traces to the Isle of Wight, VA and later to Dillon, SC and Robeson Counties, NC. His grandmother Clarke who died in 1918 listed her race as “Croatan”. How do I begin to connect with the Hatteras Y or other project, as mentioned here? Thank you for any help you can offer.
Aimee Knox Powers
The tribe now known at the Lumbee were known as the Croatans for quite some time, which adds confusion to all of this. That is likely what his grandmother’s death certificate is referencing.
You and I met at the last annual conference in Houston, Texas, November, 2013.
I am the Group Administrator for the Wesorts-Piscataway research group.
To make it brief,…I’m writing to address two issues: (#1)… because I am a bit baffled by the ethnicity results that I received from FTDNA.
(#2)..to request that you become the Co-Coordinator for the Wesorts-Piscataway project.
Issue #1 Years ago,…prior to joining FTDNA, I had done a DNA, 2.5 test with GeneTree, and the GeneTree results are quite a bit different from the FTDNA results.
After attending the last confrence (where I met you) in Houston, Texas,….among other things I learned of how to calculate one’s genealogy using the mathematical method.
According to the mathematical method,….my European ancestry on my paternal side, is definitely 25%. (= a fact that I’ve known since childhood, because very early-on my mother had told me that my own father was “mulatto” due to the fact that his father,..in her words…was “a white man from Spartanburg, South Carolina””. Therefore FTDNA’s results showing 25%-European doesn’t surprise me at all,…however, the other part of FTDNA’s results did not agree with GeneTree’s results. In contrast to FTDNA, GeneTree’s results seems to allow for the PROCTOR-side of my Wesort-Piscataway paternal ancestry, as well as the breakdown of my mother’s Creole ancestry from British Honduras (a place which is now called Belize). I paid more than $250. for the GeneTree test, and I still have the CD diskette that was sent to me from GeneTree, giving my ethnicity as : (1) 31% European. (2) 58% Sub-Saharan African. (3) 5% – to – 11% = Not sure, with margin for error. May I please hear from you as soon as possible about this issue, #1. ************************************************************************************************************************ Issue # 2: After receiving the FTDNA “Family Finder” results, I saw that we are “matches” with a 2nd to 3rd-cousin range. Therefore,…I would like to request that you become the co-Admininstrataor for my group. May I also, please have your prompt response regarding this request. Thank you. Yours truly, Carmen Proctor-Cook
Wait a minute: remember how we have two family trees, a genealogical one and a genetic one? Doesn’t that invalidate comparing your genealogically derived ethnic percentages with your genetic percentages? You could have 100% genetic contribution from that 14.5511% documented Dutch heritage, meaning 14.5511% of your DNA is of Dutch origin. Or you could have none of it, or something in between.
You’re right in that we don’t get 50% in each generation, but it would be extremely unusual to get either all or nothing, and the average is indeed, 50%. However, you can see my study on generational inheritance here: http://dna-explained.com/2014/02/19/generational-inheritance/
Hi Roberta. Yep, I had your February post in mind. The problem with the average of 50% is that we’re finding the probability distribution isn’t even (see your own Expected Inheritance Frequency and Actual Inheritance Frequency plots). It’s more like an inverse bell curve: we’re more likely to get either very little or most of a segment than half of it. When we’re talking about large percentages of our total genome that’s not much of an issue, but for very small admixtures of just a few percent, we’re dealing with a handful of small segments and so the “jitter” of the almost-all-or-almost-nothing probability could throw a wrench in things.
For example, I think it throws off FTDNA’s FamilyFinder relationship estimates for cousins. Among my matches for whom I’ve traced a connecting paper trail, I’m finding that most are 2-3 generations more distant than FamilyFinder predicts (3-5th cousin predictions turn out to be 8th cousins, etc). I attribute this to “sticky” segments from our common ancestor that had a good run of being passed down mostly intact. Then there are the cases of documented cousins not reaching the minimum threshold for a match in FF. I really think you were onto something in February, but I think the ramifications of 50% actually being rare go further than what you covered at the time.
I’d like to see a much more exhaustive study. I think that’s beyond what citizen scientists can reasonably do though.
Family Tree DNA posted an article today by Razib Khan, one of the developers of the myOrigins software that explains about the goals of myOrigins – and the results. https://www.familytreedna.com/learn/ftdna/behind-myorigins/
The new FTDNA breakdown of my ethnicity is puzzling to me. I have a Swedish grandfather, which accounts for some of the European northlands 39% and a northern Italian grandmother which accounts for some of the European coastal plain 43%. My mother’s family, as far back as it can be traced, is substantially British. Going back, say, eight or nine generations, virtually all the surnames of her ancestors are English, Scottish, Welsh, or Irish. Yet, the breakdown of European ancestry gives me European coastal island 4%.
As a medieval historian, I well know the movements of people across time, but this still does not make sense to me. It would help to know, say, whether, the “my origins” map meant to represent where our genes were 1000 or 3000 or 6000 years ago. Yet the broader description of each of the ethnic categories takes each area through time to the present. Around 1600 something like 90% of my mother’s ancestors were from those two islands, which would give me something like 45% coastal island.
I checked with Dr. David Mittelman on the question of how far back in time the results cover, and he provided the following information:
Anatolia & Caucasus – On the order of 5-10 K
Asian Northeast – On the other of 5-10 K
Bering Expansion – On the order of 10-15 K
East Africa Pastoralist – On the order of 3-5 K
East Asian Coastal Islands – On the order of 5-10K
Eastern Afroasiatic – On the order of 5-10 K
Eurasian Heartland – On the order of 5-10 K
European Coastal Islands – On the order of 2-4 K
European Coastal Plain – On the order of 1-3 K
European Northlands – On the order of 3-5 K
Indian Tectonic – On the order of 3-5 K
Jewish Diaspora – On the order of 1-2 K
Kalahari Basin – On the order of 50 K
Niger-Congo Genesis – On the order to 2-4 K
North African Coastlands – On the order of 5-10 K
North Circumpolar – On the order of 10 K
North Mediterranean – On the order of 5-10 K
Trans-Ural Peneplain – On the order of 2-4 K
This is helpful and clears things up considerably. Thanks for finding and posting the information. It would not hurt FTDNA to post it on their own website.
The main problem for all the british isles admixture results is due to the lack of detail in all tests to determine where to put the alpine people. The south germans, north italians, alpine french, swiss, austrians etc. ……..They do not even know if they are north European or south European as they all are north of 45 degrees latitute.
But the MAIN POINT is that the alpine area is home and origin of the celts and we know where they eventually went to.
I think the Family Tree DNA’s myOrigins is definitely an improvement over Population Finder although they missed some of my ancestry.
An example of my ancestry proportions from Eurogenes Genetic Ancestry Detective:
My ancestry from myOrigins:
European Northlands 51% compared to 46.9% Northwest_European above
European Coastal Plain 39% compared to 37% Central_European above
Middle Eastern 11% compared to 11% Mediterranean above
Roberta it’s great that you did this comparison. The latest Dodecad calculator by Dienekes is called globe13 and is available for download at http://www.y-str.org/2013/07/diy-dodecad-21-wrapper.html You can compare the results to the source populations with the spreadsheet found at http://dodecad.blogspot.com/2012/10/globe13-calculator.html which will allow you to see how much each country has of what percentage of each component.
As far as Eurogenes K13 components you don’t necessarily have to guess because you can see which populations have more of what component by downloading the spreadsheet at https://docs.google.com/spreadsheet/ccc?key=0Ato3EYTdM8lQdEUtZjRwTkQxRzBCeHdTaTdWUUY4Z0E#gid=0
Pingback: 10 Things to Do With Your DNAPrint, renamed AncestrybyDNA, Test | DNAeXplained – Genetic Genealogy
In reading this I noticed that there were push pins in various colors under the “Second Generation Report Card” heading. Is it unusual to only have “Exact” mtDNA matches identified on the map? I only have red pins.
It depends on if you have taken the full sequence test and if you have matches at that level, and if they entered the geographic information of their oldest ancestor.
Pingback: Ethnicity Testing and Results | DNAeXplained – Genetic Genealogy
Pingback: DNAeXplain Archives – Introductory DNA | DNAeXplained – Genetic Genealogy
It is always so interesting to read your articles. It occurs to me that the Iberian / Med mix in your DNA may be correct not because of an ancient lineage, as you speculate, but because the Spanish ruled over the Netherlands for almost 150 years between 1581-1714, and the occupation was at times truly barbaric. I, myself, have ancestry that can only be explained through wars and occupying troops. Such is the history of Europe, and sadly, much of the world.
The lack of consistency is a really a case of trying to assign a ethnicity as a recent thing of the past when it took hundreds of thousands and even up to a million years to create the 3 major human races. So in essence if it took 600K years for the caucasoid, mongoloid, and negroid races to evolve you are not going to have an easy time of classifying people based on technological and political differences to the more than 200 different countries that exist today as many of the countries aren’t even as old as we ourselves are.
I feel the reason you are getting 10% Scandinavian is because your continental contribution is being over-estimated, and that mystery 25% is almost certain to be British Isles which would bring up your total Scandinavian contribution to the numbers you are seeing. Particularly Scots surnames.
For me, having Scandinavian ancestry that is rooted in the distant Irish and British past is like my Amerindian ancestry – it’s never going to be proven with a paper trail as one doesn’t exist for most of those people. Near 100% literacy is a recent thing and to throw out all oral history because of that is short-sighted especially when so much of the oral history is being proved correctly in the more general sense of what that oral history is conveying.
So then, my mtDNA origin is also a surprise, but in it originating in colonial North America, that is the surprise. However it is not a Native American haplogroup, so if I remove the insertion that caused my haplogroup, my myDNA seems to have originated with native Finns as it is present in 12 of 18 matches on mitoSearch, with the other 6 as so: 3 being Swedes, 2 Scots, and 1 German. So it seems related to Finns learning new technology and then having a substantial population increase and it seems to have happened before Viking times. So despite the stability we are told of in mtDNA, mutations do happen and sometimes in the not so distant past. There is a small chance that my mtDNA haplogroup that I think originated in the Americas after 1650 instead originated in Switzerland in a woman whose descendant migrated to America with the Mennonites (later Amish) as one Swiss man in a large city matches us 5 Americans but many Swiss have a foreign parent and many more a foreign grandparent or great grandparent. However, it seems I’d need to contact him outside the email address he left at FTDNA and I am hesitant to do that although I can speak German well enough I’m not sure if he forgot about this obscure email server he uses that he created an account on or if he’s lost interest in genealogy.
Myself: on FTDNA I come up with 62% British Isles, 29% Scandinavian, 4% Finnish, and 5% Anatolia. And those percentages read also as an ethnic confirmation of the deep history of the British Isles so I find that interesting even if the histories are often upsetting to read.
5% Anatolia for example is it?:
a) likely the result of the PIE expansion (closely related to R1B) that created just about every ethnic group in Europe today although genetically not contributing too big an alteration to the existing populations.
b) more recent immigrants (they’d still need to be in the British Isles prior to 1740) setting up trading establishments in British port cities from Anatolia? Much less likely but being involved in international trade does give one more visibility than say a sustenance farmer.
c) Some other historical event(s)??? – the need for modern, considerate, and concise history books that use this new scientific data to confirm or refute history is needed. We don’t need to hear about the glory of empires past and present but of the technological advances and their association with those empires or spread of various languages, religions, ethnicities, and so on. It wasn’t the people that led to empire but the technology that led to empire. The technology and it’s impact outlasts the people. Technology should free and comfort, not subjugate or threaten. We need to be cautious of technology and what it is used for. We could learn a lot by learning that more technology isn’t always the answer to a problem.
4% Finnish and Northern Siberian seems to point to Arctic and Northern Siberian Ancestry in Europe and in Native Americans – I have Native American ancestry in more than one line but the ancestry precedes 1800. I was told of 2 via oral history and the tribes were identified as Shawnee and ‘injun’ but it turns out I have discovered likely 2 others: one via conventional paper genealogy and the other via mtDNA maternal research, paper genealogy, and general history books (that one I think might be Lenape). The likelihood that all of these leads can be wrong is very small. However, except for using the DNA to prove these things I grow tired of researching such lines that I know almost surely won’t have a paper trail. I have interest in the specific tribes though not out of pride but factually and respectfully. Amerindians did not have census or written language. I will better spend my time cleaning up and correcting my written tree and trying to find photographs of people and places for future generations. And I’ve found some smiling pictures in the 1800s of my Great Grandfather who actually lived to be old enough that I remember him. So those things and going back and citing references now. I am very uneasy with a reference being an online ‘Ancestry’ source.
29% Scandinavian – well the Vikings are well known but I’m willing to bet some of that is Saxon, Dutch, and Frisian migrations. Also a lot of that is likely Scots ancestors as I have several surnames that are Scots.
62% British Isles – Well from a strictly ethnic point of view just who are the ‘British Isles’ people that I am 62% ethically? I think that would be the ‘native contribution’ just after the British Isles were repopulated after the glaciers melted. You could say the same though about the Scandinavians, Anatolian, and Finns too though. If not the glaciers and most of these modern groups aren’t associated with repopulation of the north after glacier melt simply because they aren’t in the north, than you’d try to identify the technology or culture change that led to their geographic distribution.
So if I were to reclassify my genetic ancestry to pre-PIE terms and ignore the small Mongoloid / Amerindian component then I would look at the Dodecad K12b and what they seem to say is that the Gedrosian were pre-PIE people whose genetic legacy is present in the North Atlantic, Mediterranean Atlantic, and Caucasus European DNA that preceded the PIE expansion. The PIE expansion seems to have commenced from a subset of Caucasus component of my ancestral DNA.
Unfortunately reading the questions and comments on the link to the page for the Dodecad K12b project and the fact that the Dodecad K12b authors themselves make no effort to correlate their ethnical classification system with modern ethnicities in a easy to read way for those interested in the Dodecad K12b work. Genealogists with interest of only the recent past will have little interest in Dodecad K12b but those interested in the more ancient folk comprised of Neaderthals, Denisovans, sub-Saharan Africans and the various pre-PIE cultures that existed from these people will have interest in Dodecad K12b.
So to summarize a long winded point briefly – the genetic ethnic composition charts need association with technological advances, languages, ethnic groups, religions, and empires along with date ranges for something like those things to make much sense to the average layperson hobbyist that looks at a modern globe and sees over 200 countries. I don’t think any of the genetic genealogy sites do a good job of that yet.
I will definitely be using dna-explained to improve the accuracy of my written tree so thanks.
Pingback: Ethnicity Testing – A Conundrum | DNAeXplained – Genetic Genealogy
Pingback: DNA testing cannot confirm Native American Ancestry: PSA | AmandaPandaDUH
Hi, Miss R Estes
About 5 days ago i received my FTDNA Family Finder Results – MyOrigins, the results were i have 100% East Asian: 65% Southeast Asian + 35% Northeast Asian. My Geno 2.0 version results were 65% Northeast Asian + 35% Southeast Asian. I also took both my cousins DNA and their results were: Clarissa have 100% East Asian (59% Southeast Asian + 41% Northeast Asian) and Grant also have 100% East Asian (56% Southeast Asian + 44% Northeast Asian) – Geno 2.0 NG Versions: 56% Southeast Asian and Oceanian + 44% Eastern Asian. I saw a lot of similarities about the Geno 2.0 NG and FTDNA Family Finder – MyOrigins World Regions result. Was the Geno 2.0 NG and FTDNA Family Finder used a similar, or exacly same method, etc….. to determined their participants Autosomal DNA?
I see some Estes related by gedmatch.com
I am share German origins and have similar results. T898920
My Geno 2.0 Autosomal Ethnic Percentages: 65% Northeast Asian + 35% Southeast Asian, 1st Chinese and 2nd Kinh Vietnamese. My FTDNA Family Finder MyOrigins: 65% Southeast Asia + 35% Northeast Asia = 100% East Asian.
My Cousin’s Grantrussel Geno 2.0 NG (Non Helix Version) and His FTDNA FF MyOrigins results: 56% Southeast Asia + 44% Northeast Asia = 100% East Asian. His Geno 2.0 NG and FF (transferred from His Geno 2.0 NG) was exacly same! His Geno 2.0 NG Closest Ethnic Groups: 1st Filipino and 2nd Bougainville Nasioi Oceanian.
Do The Genographic Project Geno 2.0 NG use a same tool,…….etc with FTDNA Family Finder (MyOrigins)?
You’d have to check with them for the answer to that question.
You are correct that these tests don’t follow country boundaries – some of those tests are picking up deep ancestry, while others don’t necessarily follow the map boundaries they claim. For instance, Eupedia, made their own 23andme maps based on the database they had and according to them the french/german component actually follows along the Rhine River and the Scandinavian component peaks in a small area between Norway and Sweden, but is much weaker in Denmark. The British/Irish component actually follows areas with high Celtic ancestry such as Ireland but isn’t quite as strong in parts of England where the Anglo-Saxons settled. The Northwest European component appears to follow areas of Germanic settlement and peaks in Germany, Denmark, and the Netherlands, but is also very strong in the Anglo-Saxon settlement areas of England. I have search online for those who actually live and have deep roots in various countries who have taken 23andme tests. A few who have deep ancestry in England for instance show no British/Irish ancestry at all but do show a very high Broadly Northwest European component. The tests show what they show and aren’t necessarily wrong provided the results are interpreted correctly. In that case I would say that English person inherited whatever markers they test for from their Anglo-Saxon ancestors and not so much from the Celtic side.
My wife and motherinlaw are 100% German. Their 23andme results are:
Eastern European 5.1%
Broadly Northwest European 24.9%
Broadly Southern European 1.8%
Broadly European 3.4%
Eastern European 2.4%
Broadly Northwest European 26.7%
Broadly Southern European 1.7%
Broadly European 3.0%
I am about 1/4 German, 1/4 Austrian, and the rest is mainly Scottish and English.
Broadly Northwest European 28.1%
Broadly Southern European 5.5%
Broadly European 4.2%
By the way I also live in Virginia now and work with somebody names Estes. I wonder if you ar related?