[One quick note: As always, we receive no financial benefit or consideration for any product or service we review/recommend/discuss here. Everything we discuss is our opinion alone, and we talk about it because we use it.]
Ancestry has made a lot of noise recently when they updated their Ethnicity estimates, and the now intensified debate about the “accuracy of DNA tests” and the confusion among the general public makes it clear: as a community of serious researchers, we need to be the voice of reason when it comes genetic admixture and call it out for dubiously valuable, largely inaccurate parlor trick that it is. Here’s why:
Ethnicity cannot be tested for. Ever.
Ethnicity is a social construct. Period. If we look at any test, any genealogical tree or other determination it will not build a social link to ones ancestral background. Michael hasn’t been to Ireland, but I have, and despite being able to trace 12.5% of my 3x great grandparents to Ireland, and Ancestry’s admixture pointing to an Irish background, I am not Irish. I visited Ireland as an American…a very obvious American. As will Michael when he visits. Nor will he be mistaken for Beninian when we visit Benin. We are Americans, some with European ancestors some with African ancestors as well, but even with a perfect admixture that could pinpoint our ethnic ancestors exactly…we’re still not German, or Cameroonian, or English/Irish, etc. You can’t test for it, and DNA gives you no indication of how someone identifies ethnically. And that’s important, because Ethnicity is only about how someone identifies themselves and/or how others identify them…it’s not based on a gene. Neither is race, but that’s another rant for another day.
We need to voice a supportable, honest, accurate narrative to drive continued testing…one that will continue after the “Ethnicity” emperor is shown to have no clothes.
It’s not honest
All DNA testing companies, especially 23andMe and Ancestry, are for-profit enterprises that have a strong incentive to grow their number of DNA tests. The larger the test database, the more money the companies charge to sell access to your data. This isn’t to say they are selling personally identifiable data, the data is largely de-identified and aggregated, but it’s YOUR data…and it’s very, very valuable. 23andMe survives almost entirely on the revenue generated from your data, and it’s likely Ancestry is generating a large amount of their revenue from your DNA data as well. And no one’s advertising “come test with us, we are selling to great causes like Michael J. Fox Foundation” [23andMe], they are basing their sales pitch on the shiny bauble that gets the tests in the door: Ethnicity and pretty graphs. The more we play into the Ethnicity debate
It’s not our tool
Ethnicity (as determined by genetic admixture), has almost no genealogical or family history value, and the results will never break a brick wall or significantly add to your family’s stories. First, all of the major providers target who your genetic ancestors were 800-1000 years ago. Even those of us with great trees rarely go back to 1000-1200 AD…and we doubt there would be much value in anyone researching our 28th great grandparents. We have over 1 million 18th GGP’s. Admixture doesn’t rank even among the top 20 tools we use to build our trees, and it doesn’t deliver us any value.
It’s not accurate, and it’s not scientific
The biggest red flag from Ancestry’s last update was this: they increased the reference samples from 3,000 tests to 16,000. They have literally spent the last 4 years selling “Ethnicity” to the general public as a great reason to build Ancestry’s test database, even though the entire house of cards was built on 3,000 reference samples. There is no statistically valid data that be gleaned from 3,000 total samples as they relate to our genetic ancestors 1000 years ago. Again, we each had MILLIONS of ancestors 30 generations ago…and to use 3,000 for all genetic admixture just demonstrates the shoddy science that underpins this process. Even 16,000 is a ridiculously small sample…even if they were each perfectly tied to a region 1000 years ago. “Ethnicity” is just enough science to seem valid enough to be scientific…and just scientific enough to justify the pretty graphs that facilitate the selling of more tests.
It’s hurting genealogy, and it will ultimately turn the public off of genetic DNA testing
Youtube is rife with videos of the general public discussing their “inaccurate” DNA tests, with the testee going into great detail about how they know their Ethnicity and they see something they don’t expect, the test is wrong. There are now new discussions everywhere with people questioning the entire testing process when the “results” can be changed so dramatically by a change by Ancestry. Ancestry is aware of the strain this update is having on the general public, and we can see the efforts they’re making to try and calm people as they go through the update. There are explanations, surveys, etc. to try and make sure the public doesn’t freak out about this change. It’s all just adding more weight to the idea that these tests aren’t accurate/reliable. Since the entire business case for the public taking these tests has been “Ethnicity”, once that’s being exposed as the subjective “art” that it is, the only reason for people to test is being questioned. We will hit a tipping point where our relatives are going to think of DNA testing as a “scam” that’s of no value/dangerous, and it’s going to make the process of getting tests that much harder.
So, what can we do? What impact can we have? Honestly, not much…at least not immediately. But, as the people serious about genealogy we can start being the voice of reason and begin to lay out a better justification for why the public should test, even if the focus of the commercial testing companies is only on adding more samples to their databases. If the thought-leaders and respected voices in the communities turn their back on genetic admixture, that will eventually drive the discussion.
To that end, here’s our suggestions:
- Stop discussing “Ethnicity” as a testable value – Push back on this basic premise and start to educate the public on why DNA tests have no value as it relates to how they identify ethnically.
- Don’t give genetic admixture a place at the table – We should no more engage in admixture as a point of genealogical value as we phrenology. They both sound scientific, and their proponents would like them to be seen as science, but neither are science. Even making an anti-admixture discussion elevates it to a “con” in a pro vs. con debate. We need to stop engaging in a debate of equal positions with admixture.
- Develop other reasons the general public, and our relatives, should submit tests – The tens of millions of tests in various databases have a HUGE value to the genealogical community, and we all benefit as more tests are added. We need to voice a supportable, honest, accurate narrative to drive continued testing…one that will continue after the “Ethnicity” emperor is shown to have no clothes.
- Be honest with our relatives as they test and help them, and the general public, understand how these tests play into the for-profit world – Those who take tests aren’t purchasing a product, they are the product. 23andMe and Ancestry needs those tests to make a profit, and it’s the only reason why they offer these tests. Let’s discuss that, and what we get in return, to level set everyone’s expectations. If we don’t set these expectations, some scandal will do it for us, and when negative public opinion sets in, we likely will lose the value of having non-experts testing.
Bottom line is that we can see how the reality of DNA testing doesn’t match the perception of the testing public, and all eggs are in the “Ethnicity” basket. As that basket starts to fray, we can either be a part of the rational message that keeps this testing world moving forward, or we can be reactive and wish we could go back to the “good old days” when people were testing without us having to fight for each one.
13 thoughts on “It’s time to stop giving attention to “Ethnicity” and genetic admixture”
Reblogged this on The Family Kalamazoo and commented:
This blog post responds to the ethnicity estimates on Ancestry as did my last blog post. This one is more eloquent. 🙂
Well, I enjoyed your writing on this as well…but thanks for the resharing! 🙂
Sure. And thank you!
I agree 100%, and every time a friend (who otherwise doesn’t do genealogy) asks about DNA testing, I tell them this as well. It’s not science, it’s just statistics, and it is a waste of money.
Amy, thanks for the comment…and you highlight exactly what worries us about selling DNA tests to the public based on “Ethnicity” reports. DNA tests are both HIGHLY accurate, and very, very valuable when establishing one’s genealogy. Especially in African-American genealogy, even when we don’t have complete trees to compare, narrowing down where to search or where possible matches might connect is invaluable. We know we have brick walls that will only be broken down because of our DNA tests.
But, because these tests are sold to the public based on “ethnicity”, and the genetic admixture that underlies these tests is so flawed, people will eventually equate the accuracy of the test for genealogical purposes to the the inaccuracy of the tests for “ethnicity”, and will decide not to test because they’re not accurate.
They are accurate, but really only for doing your family history…and I’d love for you to test and add to the data as we try and build out as many of our ancestors as we can…but to do so knowing that’s why you’re doing it 🙂
Oh, I have tested as have several of my relatives. As an Ashkenazi Jew, however, I have thousands of matches, most of whom are not ever going to be traced as actual relatives, so it is pretty useless for me even for genealogy purposes. But I know that for many others, it has been incredibly helpful.
So luring people into testing with the misleading claims about ethnicity can be a good thing since it increases the data pool, BUT unfortunately most of those people get their results, do no other family research, do not respond to inquiries, and thus only frustrate those of us who are doing serious genealogy!
Thanks for your response!!
Thanks for reposting this Luanne.
Great blog post. I agree with much of what you say. However ethnicity can help at times in your family research. I do agree that people can spend too much time and attention to these findings. But myself I find them fun and shows in a real way how all people in this tired old world are closer than many think. If these test opens just a few more “closed” minds about people they are well worth it. Besides I truly feel as time goes on they will get much better.
I have tested with three companies, and have very different results from all. AncestryDNA’s new result is totally different from the previous result, and is broad enough to be very accurate. So I definitely get that these results are to be taken with a grain of salt. But I do know people whose ancestors were from one place, and changed their names without telling their future children. DNA testing revealed where they were actually from, and this was a great revelation for the family. They were then able to do the genealogy work to find the connections. So while not what it is being sold as, still not totally useless. But the cousin matching is really where it’s at.
DNA is useful for African Americans who don’t ‘know’ which country they came from- as I ‘know’ I’m from Italy/Ireland etc… or even who a 2nd or third great grandparent was… DNA has open a firmly shut door to lots of research avenues…. we are thankful…
That’s a great point…and as someone who’s tree if 1/2 African-American, there is a huge value in knowing any firm fact given how few facts can be validated compared to European ancestry.
The point we were making was that knowledge doesn’t translate to genealogical/research value since it’s focusing on who our ancestors were 1000 years ago…which would encompass over 4 million people (give or take) and it won’t be able to reliably help us match to other DNA testers or to build out our trees. In fact, the hope that we can use their “Ethnicity” estimates to narrow down our DNA matches is just a sign of how poor the tools on Ancestry are! There is a large selection of Admixture tools on GEDmatch…which we’ve never used, because we have like 8 really valuable DNA tools to use to narrow our matches.
The other challenge is that, just like in many things, African populations are underserved in Ancestry’s reference samples for their admixture calculations. When we tested our African American ancestors, there were a TOTAL of 3,000 reference samples…only a small portion of which were African. Even now, as they expanded to 16,000 total samples, the number of African samples is likely much less than European. Native American populations are even more impacted. What it means is that the important connection to where home was isn’t based on good solid science, and we’re seeing our relatives we tested quite bothered by the sudden change. To build this connection to Nigeria because the admixture originally said that a relative was 38% Nigerian, only to have the new test say she’s 3% Nigerian (which could actually be 0-15%) means we’ve taken that one thing of value for African Americans in “Ethnicity” and shown them they can’t trust it. Our relatives are upset, and we won’t get anymore tests from them for at least awhile.
We’ve been looking at a trip to Africa through Roots to Glory (http://www.rootstoglory.com/) and if we had gone to Nigeria because we thought my wife was 38% Nigerian, built a connection to what we thought was home, and then found out later she’s essentially not Nigerian…it would be painful.
The risk of using admixture/ethnicity to sell these tests is that these changes will turn people off of testing. That testing has been SO key to building out parts of our tree where people did get recorded by the majority system, if people stop testing because of admixture, it’s going to set us all back.