DOES A DNA TEST REVEAL YOUR ORIGINS?

I really don’t like misleading TV ads. And there is no shortage of those around. One ad from Ancestry.com in particular has bugged me for some time. It goes more or less as follows:

The not-so-subtle implication is: you may think that you have descended from one particular nationality, when in actual fact you have descended from a completely different nationality. And a DNA test will provide scientific and incontrovertible evidence for your genetic heritage. So take our DNA test. That’s the implied message of that ad.

That whole scenario is so utterly stupid!

Why do you think that they grew up "thinking they were German" in the first place? It’s obviously not because great-granddad John McHenry had come to Ellis Island from Aberdeen, Scotland, is it? No, they grew up "thinking they were German" because great-granddad Johannes Heinrich had come to Ellis Island from Hannover, Germany.

The reason people grow up "thinking they are of a particular national background" is because they know the identity of one or more grandparents or great-grandparents who came from some other country. It is ridiculous to think that people claim one specific line of descent as a major chunk of their overall ethnicity, if they don’t know the identity of a single ancestor within the past 100 years that represented that particular line of descent.

People think that they have a German / French / Belgian / Scottish / Dutch / Danish / Swedish / Norwegian / Russian / Spanish / Irish / Greek, etc. heritage because they know about one or more ancestors who came here to the USA from Germany / France / Belgium / Scotland / Holland / Denmark / Sweden / Norway / Russia / Spain / Ireland / Greece, etc.

What the TV ad implies is: you can’t trust that knowledge about one or more particular ancestors from the old world. No, you need our DNA test to tell you what your background really is! And you can’t argue with science! Your ethnic background can only be established with certainty if you take our DNA test.

So I wanted to find out for myself just how accurate are the results from such tests. Therefore I decided to take a DNA test myself.

Now in scientific research circles test results from a study involving people’s traits, characteristics and identities are only considered reliable when they are obtained from a "blind" or a "double blind" study, as the case may be. By this researchers mean that the people who might in any way influence the evaluation of the research results that are obtained must not know the identities of the people whose data they are evaluating. This is to avoid the temptation for researchers to make their results conform to the expectations or to the desires of those commissioning the research, a practice that is unfortunately not at all uncommon.

Let’s apply this point to DNA research:

That is so because last names openly reveal the nations of origin for those names, before any DNA samples have even been looked at.

So I wanted to have my DNA evaluated in the equivalent of "a blind study". In other words, if my DNA really does reveal my racial background, then that should not be influenced by whether my name is Irish or Russian or Spanish or Chinese or Indian or Italian or any other nationality. The DNA is supposed to tell the story totally independently of whatever my name might happen to be. A name has no influence on someone’s DNA, right? After all, how many ex-slaves took on names like"Washington" and "Jefferson", etc., without such names having any influence on their DNA?

So to achieve "a blind evaluation of my DNA" I did the following:

1) I gave a false name from a totally different nationality to my own ethnic background. I am German, and both of my parents were Germans, and all four of my grandparents were Germans. So I had Ancestry.com do my DNA test for the last name of "MacSomebody" (I will keep the actual name I used confidential), together with a first name that would likewise suggest that I might be either of Scottish or Irish descent. (Perhaps I could receive a result that would also allow me to trade in my Lederhosen for a kilt?) Using an incorrect name or nationality should not have any effect on my actual DNA profile, right?

By way of explanation: once I had chosen my false "Mac..." name, I then did an internet search on that name. Once I knew that hundreds of other people actually have both the names I had selected, then I stayed with that name. If the names I had chosen would have turned out to be very rare, then I would have changed one or both names, to achieve a name that turns up scores or even hundreds of results in a search. That would have been to preserve as far as possible the anonymity of my real ethnic identity.

2) I used an email address that does not contain any part of my real name. This also makes it somewhat more difficult for others to verify my real identity.

3) I used a postbox address for the mailing of the DNA sample kit. This was likewise aimed at minimizing the chance of the DNA researchers confirming my real identity by means of my residential address.

4) I never provide any correct personal data over the internet that is not absolutely required. I will use false names and false dates of birth, etc. when such information is asked for by people who I believe don’t have any right to know that kind of information about me; e.g. filling out some or other form to be granted access to some information on some webpage, etc. After all, I myself don’t ask visitors to my website for their names and dates of birth either; so I don’t see why other websites should be entitled to that type of information. I strongly resist anyone on the internet (and for that matter, also over the phone) seeking to obtain any personal information about me.

So when Ancestry.com asked for my year of birth, then I provided a wrong year. Whether I was born 5 years earlier or 10 years later will not have any effect whatsoever on my DNA profile. So a request for my year of birth from people who claim that they will evaluate my DNA was a warning flag.

In genealogical research there is a need for the correct year of birth, whereas the month and the day of the month are only of secondary importance in that type of research. So when they asked for my year of birth, I realized that Ancestry.com was going to do some major genealogical research through the billions of records that are out there in the public domain. But I don’t want to be given an ethnicity profile based on genealogical records! I want them to give me an ethnicity profile based exclusively on my DNA. After all, that’s the service the Lederhosen guy was advertising.

Concerning DNA research: for any one of us to completely take all genealogical records out of the equation requires only two things: a false name that is associated with a different nationality group, and a fabricated year of birth. And then make sure you don’t reveal your residential address. That will then result in a "blind study", one where the DNA research lab is restricted to using only the actual DNA evidence itself. The false name and the false year of birth will not provide them with any significant information that could be used "to guide" them in the interpretation of the data which they will obtain from the DNA sample. I don’t want them to be "guided" by any non-DNA information.

5) I also refused to give them permission to share my DNA results with any other database or with anyone else. And I obviously refused to start creating some kind of genealogical chart for my supposed ancestors. That can’t really be done with a fake name.

Let’s understand that the invitation to start building some kind of genealogical chart is aimed at benefitting them and not you! It will make it much easier for them to find genealogical information about you, because such a "family tree" will force you to provide to them correct data about your national and racial and family background.

Note that if you don’t start such a family tree within a week or so, you will probably receive a friendly reminder from them to start setting up your family tree. That reminder tells you that they are waiting for this voluntary genealogical information from you. Any information you provide here will guide them in interpreting your DNA.

Now nobody out there on the internet has the right to know anything about your genealogical history, and you should never divulge such personal information to anyone on the internet.

6) Where possible, try to use someone else’s credit card to pay for the DNA test, preferably someone of the opposite sex, to maintain your anonymity. (Give a friend the $69 in cash for him or her to pay for this test with his or her credit card.) A credit card payment from someone of the opposite sex and with a different last name will tell them that it will be futile to associate that friend’s name with your DNA ... should they attempt to establish such a connection.

In short, I sent them my DNA sample in such a way that they were forced to rely exclusively on what the DNA analysis would reveal. I didn’t give them any hints of any kind regarding my true racial and national background. The fact that I implied a different national background for myself should not in any way affect the actual analysis of my DNA.

So I sent my DNA sample back to them. They had cautioned me that it might take 6-8 weeks for my results to become available, due to their "heavy workload". However, a mere 20 days after I had put my DNA sample in the mail to them, they informed me that my DNA results were now available. Later I will present what I believe may be the reason why my results were available so quickly.

Okay, now let’s get down to the good stuff.

Having logged into their website, I was presented with a page that is titled "DNA Results Summary for ... (the two false names I had chosen)". Across the top of the page is a two inch wide band with a map of a strip of the Northern Hemisphere, going from North America across the Atlantic to Europe and then across to Asia (with Siberia, etc. at the other end).

On that map there are three big "circles" (not strictly circular, but reasonably close) which are shaded over the countries covered by those "circles". Those three "circles" represent the three main sources for my supposed DNA origins.

Now it becomes really interesting! For my personal results:

One "circle" covers 16 countries. Those 16 countries are: Poland, Slovakia, Czech Republic, Austria, Russia, Hungary, Slovenia, Romania, Serbia, Ukraine, Belarus, Moldova, Lithuania, Latvia, Bosnia and Herzegovina, and Croatia. (Only some of the western parts of Russia are covered in that "circle".)

One "circle" covers 3 countries. Those 3 countries are: Norway, Sweden and Denmark.

One "circle" covers 6 countries (or ethnic groups). Those 6 groups are: England, Scotland, Wales, Holland, Belgium, and northern France. This "circle" very carefully avoids including any part of Ireland.

The visual message of this strip-map is this: My DNA can supposedly be traced to over 20 different countries on all sides of Germany (except the south!), but with absolutely no connection whatsoever to Germany itself!

Below that map in rather small print is the following information:

"Ethnicity Estimate:

Europe East (Czech Republic, Slovakia, Poland & Lithuania) = 72%.

Scandinavia (Sweden, Norway & Denmark) = 16%.

Great Britain (England, Scotland & Wales) = 7%."

Those figures only add up to 95%. So I had to click on something else to get information about the remaining 5%.

They refer to this remaining 5% as "Low Confidence Regions", to tell us that this information is less certain. In my report they present those remaining 5% "low confidence regions" as follows:

"Ireland / Scotland / Wales = 3%.

Finland / Northwest Russia = 1%.

Europe South = <1%."

(The designation "low confidence regions" obviously implies that they are quite confident about the other listed regions.)

And that’s it! That’s what my DNA has supposedly revealed.

So if I limit my search to the information they have provided in the opening screen, then I am supposedly:

72% from the Czech Republic + Slovakia + Poland + Lithuania

16% from Sweden + Norway + Denmark

7% from England + Scotland + Wales

3% from Ireland + Scotland + Wales

1% from Finland + Northwest Russia.

Okay, I’ll accept the 1% as unspecified. That’s not a problem.

If I were to believe that this DNA profile is correct, then I am supposedly one quarter from the British Isles & Scandinavia. But that is simply not true!

Similarly, if I were to believe that this DNA profile is correct, then I am supposedly almost three quarters from the Czech Republic & Slovakia & Poland & Lithuania. But that is likewise simply not true!

For the 7+5 = 12 regions for German ethnicity covered in their reports the unequivocal assessment is "No Connection"! That is likewise an utterly absurd assessment!

I know for a fact that my grandparents came from Saxony (i.e. the German Saxons, not to be confused with Anglo-Saxons), and from Prussia and from the Rhineland area of Western Germany, this last area being the place where I spent most of my pre-teen years. I went to school there.

I spent 10 years living in the same house with my one set of grandparents, with very easy and regular access to my other grandmother half a mile away. From my grandparents I learned that one of the grandparents of one of them (we all have 16 different grandparents of our grandparents!) back in the early 1800's had come from "somewhere east" ... I don’t recall whether it was from Poland or from the areas now known as the Czech Republic and Slovakia, or from Russia, or from some other country east of Germany. But all four of my own grandparents were indisputably German, from Saxony, Prussia and the Rhineland, as I mentioned earlier.

So yes, there could quite possibly be up to perhaps 5% ethnicity for "Europe East" as they call it. But 72% for "Europe East" is absurd!

So the DNA results provided to me by Ancestry.com are totally wrong! They present a totally false picture! Their total omission of Germany in my DNA profile proves this.

What I have thus far presented is based on the information that they present in the opening screen. But there is also the option to dig a little deeper into this information. And that is what I did next.

Ancestry.com has divided up the world into over 350 regions. We’ll examine the 4 greater areas that supposedly make up 98% of my DNA. For our purposes let’s say that they have exactly 350 regions, to illustrate the point I wish to make.

Now for each region in the world they provide a comment regarding the degree of connection the specific DNA under consideration has to that particular region. I am not aware of all the levels of connection that Ancestry.com uses, since only two of them appear in my profile.

The two degrees of connection on the very bottom rungs of their classification ladder are:

1) No Connection

2) Possible Connection

These two levels clearly imply that there must also be some "higher levels" of connection. There surely must be more than just "possible connection", or else all their DNA testing would be pretty useless, if they could establish nothing better than "possible connection". Understand that "Possible Connection" is not a very confident assessment at all. "Possible Connection" leaves open the door that it may turn out that in actual fact there is "no connection" at all. This level is only a very vague and imprecise assessment.

So I suspect that they have at least two more "levels of connection", if not more. Those two additional levels should more or less cover the following criteria.

3) Quite Likely Connection

4) Definite Connection

Now for my personal DNA profile they present the following picture:

1) For 349 of the 350 regions in the world they give me the assessment "No Connection"! These 349 regions with no connections include all the regions in Scandinavia, for which they have given me 16%, and all the regions in Great Britain, for which they have given me 7%, and all the regions in Ireland / Scotland / Wales, for which they have given me 3%.

What this means is this:

For all of the regions for which they have given me a total of 26% (i.e. Scandinavia + Great Britain + Ireland / Scotland / Wales), when you dig down into the report, you find that they actually acknowledge that my DNA does not have any connection at all to any of those areas that supposedly provide 26% of my DNA.

So how scientific is assigning to me 26% for those areas?

In my case "No Connection" is a correct assessment for all those 349 areas, except for the regions that cover Germany. When they state that my DNA has "No Connection" to any area of Germany, then they are obviously wrong in their assessment. But here is the point:

This is what "the small print" in my report states. So how on earth can they justify assigning 26% of my DNA to areas, to which my DNA in fact has no connections at all?

But let’s continue.

The only region for which they give me a "Possible Connection" is the area for which they have given me 72%! That is the area that is comprised of the Czech Republic plus Poland plus Slovakia plus Lithuania.

They assign nearly three quarters of my DNA to four specific Eastern European countries, even though they have found nothing more definite than a "Possible Connection" to those four countries.

That is a pathetic research result!

It is in fact an acknowledgment that they don’t really know where my DNA comes from. And that is due to the fact that they couldn’t trace my "MacSomebody" name in any genealogical records anywhere in the world. It demonstrates that without access to the real name and birth year (to avoid following a trail of another person with the identical name!) they were not able to establish any ethnicity at all. All they could do is suggest some very vague and generalized "Possible Connection", which suggestion happens to be overwhelmingly wrong.

Now this is not information that you see if you only look at the opening page for the "Ethnicity Estimate". But the point to note is that they have committed almost three quarters of my DNA to an area of the world for which they could determine nothing better than "a possible connection".

They have not been able to tie down my DNA to one dominant nationality. The best they could do is divide the dominant chunk (72%) into four more or less equal parts for those four countries. But that’s not how DNA works! People simply do not have four different nationalities to equal degrees. It doesn’t work that way!

The fact that they have refused to separate the 72% into four distinct groups shows that they lacked the confidence to do so. So they lumped these four together, to avoid exposing their inability to assign percentages to each of those four countries.

Shortly I’ll explain why they waffle around in generalities like this.

When we dig down one more level in the DNA report, then we find that they provide a range for each area. That is "a percentage range for that area".

So here is some more information.

1) For the 16 countries for "Europe East" they provide an ethnicity "range" of 62% - 79%. This range applies to the four countries they have listed. So the supposed "range" is 62% - 79%; but on the opening page they have assigned 72% to this area. How did they do that? They just picked a number somewhere in that range. That’s all they did. So someone who doesn’t know any better will look at the opening page and say: well, science has established that I have 72% DNA from this specific area. But that is not correct; that percentage on the opening page is only a very general guess, which for my DNA happens to be completely wrong.

There is a considerable difference between 62% and 79%. But you would never know that, if you only look at the opening page.

2) For the 3 countries in Scandinavia they provide an ethnicity "range" of 3% - 28%. There is a huge difference between 3% and 28%. But you wouldn’t know that either from the opening page. So here they opted for a middle-of-the-road choice of 16%. Once again they are waffling around.

3) For England, Scotland and Wales they provide an ethnicity "range" of 0% -19%. Here they acknowledge that they actually have 0% ... and yet they are willing to go up to 19% in their supposed range. Why?

4) For Ireland, Scotland and Wales they provide an ethnicity "range" of 0% - 9%. So again they acknowledge that they actually have 0% ... and still they are willing to go up to 9%. Why?

We should understand that these "ethnicity ranges" were not derived from the DNA analysis! They were derived from their genealogical research. The DNA sample itself simply does not present "a range" for each different nationality.

So the final picture that Ancestry.com provides for my DNA looks as follows:

1) 4 Countries in "Europe East" = a range of 62% - 79%, and they have assigned 72% to this area. They refused to specify percentages for each of those 4 countries. Why?

2) Scandinavia = a range of 3% - 28%, and they have assigned 16% to this area. Again, they have refused to specify percentages for each of those three countries.

3) Great Britain = a range of 0% - 19%, and they have assigned 7% to this area.

4) Ireland / Scotland / Wales = a range of 0% - 9%, and they have assigned 3% to this area.

Thus far I haven’t pointed out the obvious: they have duplicated things, by first assigning 7% to England / Scotland / Wales, and then assigning an additional 3% to Ireland / Scotland / Wales. This is actually gobbledygook, since they only chose 7% from their own range of up to 19% to start with. They could far more logically have assigned 10% from that 19% range to the "Great Britain" component, instead of bringing in an additional 3% through a back door. But the idea here is to confuse the recipient of this report.

Now let’s evaluate the data that Ancestry.com has provided.

Ancestry.com has access to multiple billions of records about people in very many countries around the world. They make most of their money not from the DNA tests they perform, but from customers wanting access to genealogical research. So they encourage people to build family trees.

When they receive a DNA sample for testing, they also do searches through all the databases at their disposal. The names tell them which countries and nations the people supplying the DNA are likely to have come from.

And when people give their real names, then the searches through billions of records will in many or even in most cases find records for those people and for their close relatives. Whatever DNA results are then established, they will evaluate those results through the prism of the genealogical data they were able to uncover about the specific DNA donor.

The genealogical research they perform takes far longer than evaluating the DNA samples they receive. This is where the 6-8 weeks for results to become available enter the picture. In my case they couldn’t find anything for the name and birth year I had provided, because the name I provided is fictitious. So therefore my results were available within 20 days. Had they found any genealogical information, the results may have taken a few weeks longer, and the DNA result would have been perfectly matched to the genealogical info they had uncovered. How convenient.

Anyway, when they obtain the genealogical records about a specific person, then they make the DNA "fit" into the genealogical picture they have established.

For my fake Scottish name they simply couldn’t find any records. Nothing matched the birth year or my generalized location. But it seemed clear to them that my DNA simply had to come from someone who must have genetic connections to Great Britain, Scotland, Wales or Ireland.

The fact that they didn’t find anything in my DNA that could be linked to any part of the British Isles or to any part of Scandinavia forced them to ignore this fact! They "knew" that my Scottish name had to have come from the British Isles. And the British Isles, in turn, have an established historic connection to Scandinavia. And so they provided me with a DNA picture that reflects their name-generated prejudice.

In fact, they provided a DNA picture that would also allow me to trade in my Lederhosen for a kilt. That’s really nice of them.

They were clearly not adverse to finding non-British influences, as evidenced by the 72% they wrongly assigned to "Europe East". So when they were willing to find the Czech Republic and Poland, etc. in my DNA, then there is no reason why they should not have found the dominant German component of my DNA. The fact that they found ZERO German DNA shows that their confident conclusions are highly suspect. For my personal DNA profile they are obviously and blatantly wrong!

Finding a range from 3% - 28% for a Scandinavian influence reveals a huge unreliability factor. The difference between these two extremes is one whole quarter of a person’s DNA profile. That’s huge. It reveals how imprecise their DNA evaluations for ethnicity actually are!

That is what I had wanted to find out: are DNA results in fact tailored to the names which are submitted with the DNA samples? Or are DNA results really a reliable indicator of ethnicity, even when the tests are performed in total anonymity?

It seems pretty clear to me that in my case they chose to interpret my DNA data based on the "MacSomebody" name I had provided. The DNA results they provided for me are totally wrong! They totally and completely miss the dominant DNA region ... Germany!

I have no doubt that if with my DNA I had provided names like "Fritz Niedermaier" or "Juan Rodriguez" or "Olaf Eriksson" or "Giancarlo Pedroncelli" or "Armindo Carreiro" or "Alexander Afroudakis" or "Miklos Kovács" or "Isaac Goldberg" or "Hendrik van Tonder" or "Jonathan Chamberlain", etc., then I would have received a completely different DNA profile for each of those names. But it would in each case have been the identical DNA. (Obviously, now that they have my DNA in their records I cannot submit any further DNA samples under other names. That would immediately be picked up. And I have no desire to take any further DNA tests.)

Based on the results they provided for my DNA I believe that their "Ethnicity Estimate" is not really related to the actual DNA which donors provide. Their "Ethnicity Estimate" is influenced very heavily by the genealogical searches through billions of records they are able to perform for the name and year of birth which the customer has provided.

Now note! This is important to understand!

When a person provides their correct name and year of birth, then the "Ethnicity Estimate" may in fact be reasonably accurate! But that "accuracy" will not be based on the DNA sample which the person provided! That "accuracy" will be based primarily on the exhaustive searches they performed through billions of genealogical records, linking the correct name and year of birth to numerous other individuals through the generations. But even then significant errors cannot be ruled out.

It’s just that you will be under the impression that they derived this ethnic profile in a scientific way from your DNA sample, when in fact they derived it from the genealogical records their computers searched through. And they then interpreted your DNA to "fit in" with what their genealogical research had already found out about your background.

If you are interested in finding out about your real "ethnicity" without regard for whatever method the researchers may use, then a genealogical search through all the available records may be your best course of action. But if you want to find out what your DNA reveals about you, without looking at any other information that is available, then at this stage these DNA tests are totally unreliable ... as witnessed by them not finding any connection whatsoever to Germany in my personal DNA.

Appeals to "find out what your DNA reveals about you" are misleading, because such appeals are supposed to make you think that it is a scientific process which is 100% reliable. That’s not true ... not when they couldn’t even find 5% German DNA in my sample, and not when they provide possible ranges from 3% - 28%. Such appeals hide the fact that the process is not based on some scientific analysis, but on plain old-fashioned computer searches through billions of genealogical records. That doesn’t necessarily make it wrong; but neither did they draw that ethnicity profile out of your DNA sample.

Understanding how they establish their "Ethnicity Estimate" should also help you to understand why siblings with different last names may at times receive results that show significant ethnicity differences. This is especially the case if the two DNA samples for the siblings are processed by two completely different labs. The genealogical searches for the two different last names may sometimes lead the researchers in different directions, resulting in different DNA profiles. The differences are based on the records they searched for the two different last names. The differences are not necessarily based on ethnic differences in the respective DNA samples of the siblings.

If you are interested in proving this matter for yourself, then I suggest that you do what I have done. Take a DNA test, making sure that you don’t provide any information that could be traced back to your real identity. Then provide a first and a last name for someone of your sex, but from a different nationality. If you are a Caucasian, then choose another nation that is also Caucasian. If you are Asian, then choose another nation that is also Asian, etc. Don’t choose an identity that is radically different from your true identity. Do an internet search on the name you have chosen, to make sure that there are at least scores of people with those two names. Provide a year of birth that is within ten years of your real birth year.

And deny permission to reveal your DNA results to any other group or organization or database. And don’t try to establish some family tree for the fake name you have chosen.

If there were 100 people who have thus far never yet taken a DNA test, and who are willing to do such a DNA test with a false name for a nationality different from their own, while keeping their own real identity completely hidden, then it could be exposed on a large scale that these "ethnicity estimates", supposedly derived exclusively from a person’s DNA, are at this stage nothing more than guesswork. The people who evaluate the DNA samples for ethnicity should under no circumstances be aware of the nationalities of the people whose DNA they are evaluating. They must not know the real names of the people involved. Knowing the real names of the people involved is cheating, as far as evaluating DNA for ethnicity is concerned.

Remember that you can only take such a test once. After that they’ll wise up that one name must be false, because two different people can’t have the identical DNA. There is a lot that DNA research can tell us. But accurate ethnicity is at this stage not one of those things. If you are interested in establishing your ethnicity profile, then you are better off if you have them do a genealogical search for you, without any DNA being involved.

And the guy in the Ancestry.com TV ad might want to look into getting back his Lederhosen and ditching his kilt.

Frank W Nelte