DNA testing and reliability
Reader Wilma Bittinger was dismayed when her daughter sent her a link to an article on DNA testing that was published online in 2013.
Originally titled “DNA Ancestry Tests Are Meaningless for Your Genealogy Search,” the article in the online Medical Daily begins by suggesting that:
Genealogy tracking has become big business, with many companies charging up to $300 to trace your DNA to specific historical figures or ethnic groups in the distant past by analyzing ancestry tests.
A group of scientists now offers a public warning that these ancestry tests have little scientific backing, and are often so unreliable and inaccurate that they amount to “genetic astrology.”1
“Would love to know what you think of their opinion,” Wilma wrote. “I’m not trying to tie myself to any one specific historic figure but they compare DNA testing at our level to astrology. Say it isn’t so!!!”
It isn’t so.
That wasn’t what the scientists who wrote the piece the article commented on said.
And even the article itself ends up saying it isn’t so.
The scientific piece that Medical Daily commented on was “Sense About Genetic Ancestry Testing,” written by Professors David Balding and Mark Thomas, and Tabitha Innocent, Sense About Science; with assistance from Dr Turi King, Dr Lounès Chikhi, Dr Rosalind Harding, Professor Mark Jobling & Professor Guido Barbujani.2
It should tell you something to see Dr. Turi King in the list of contributors: she was one of the lead scientists who researched the Richard III case recently — using DNA evidence as part of the case that the skeleton found under a Leicester, England, parking lot was the last Plantagenet king.
So let’s begin by seeing just what it is that had the scientists upset.
What they criticized — the specific aspect that they called little more than “genetic astrology” — was the selling of DNA tests that tell you if you’re related to a distant historical figure, or that characterize your results with broad labels like “Viking” or “Zulu” or that tell you that “your ancestors moved to a particular part of the world at a specific time.”
And genetic genealogists are as critical of those advertising techniques as the geneticists themselves are.
The problem with saying you’re related to some distant historical figure is that everybody alive today traces back to the same set of far distant genetic ancestors. As the scientists put it, “If you are told that you are genetically related (share a genetic marker) to someone who lived a long time in the past, it may well be true but is not very meaningful. In reality, we all share the vast majority of our DNA through remote common ancestors.”3 That’s why you don’t find genetic genealogists telling you to add, say, Genghis Khan to your online family tree even if you’re an exact match to his known genetic markers.
As for the labels, the scientists note that “people’s genetics don’t reflect discrete groups. Even strong cultural boundaries, such as between the Germanic and Romance language groups in Europe, do not have very noticeable genetic differences.”4 So pigeon-holing results into broad categories like “Viking” or “Zulu” is patently ridiculous.
And genetic genealogists sing the same song on that issue. The Legal Genealogist has written time and time and time again about the simple reality that the percentages reported in the ethnicity estimates by DNA testing companies are just that: estimates, and nothing more.5 They are, to be blunt, cocktail party conversation starters and not a reason to do DNA testing at all.
As for the migration data, the scientists note two problems: the fact that a particular genetic group is found most commonly in one part of the world today doesn’t mean it wasn’t introduced to that part of the world more recently than we’d expect; and the fact that even if it is more common in one part of the world, you individually may have inherited it from an ancestor from another part of the world altogether.6
Yup. And genetic genealogists won’t tell you to add “North African” to your online family tree just because you and many millions of other people share that genetic marker, either.
So… so far… genetic genealogists and the geneticists are in complete agreement. There are some ways that DNA tests are being sold that are just plain silly when it comes to being useful for genealogy.
Does that mean that all DNA testing for genealogy is “genetic astrology”?
It isn’t so.
That wasn’t what the scientists said.
And even the article itself ends up saying it isn’t so.
There’s a highlighted section of the scientists’ report that reads as follows:
There are some things genetic ancestry tests can tell you quite accurately.
There are credible ways to use the genetic data from mtDNA or Y chromosomes in individual ancestry testing, such as to supplement independent, historical studies of genealogy. If, for example, two men have identified – through historical research, possibly involving surnames – a common male-line ancestor in the sixteenth century, it would be reasonable to use their Y chromosome data to test this. There are some ancestry testing companies that offer this service.
To answer a specific question about individual ancestry with any degree of confidence requires a combination of historical records and genetic information.7
And boy do genetic genealogists ever agree with that: DNA testing only works with traditional paper trail genealogy, not instead of it! As much as we might wish that DNA came with each segment neatly labeled with the name, birthdate and birthplace of each ancestor who passed it down to us, it just doesn’t work that way. Using DNA in genealogical research is hard work; it’s part of our proofs, part of our evidence, and takes a lot of effort to use it right.
But because DNA combined with historical research is such a powerful tool, even the article ended up backpedaling away from its original criticism. Medical Daily changed the title of the article from “DNA Ancestry Tests Are Meaningless for Your Genealogy Search” to “DNA Ancestry Tests Are ‘Meaningless’ for Your Historical Genealogy Search” and it added a note at the bottom of the article:
The original version of this article was entitled “DNA Ancestry Tests Are Meaningless for Your Genealogy Search,” which was inaccurate. The source material does not question the usefulness of DNA testing for questions about immediate biological relations, like paternity tests or adoptees looking for their biological families; its criticism was limited to DNA ancestry tests that claim to answer specific questions about ancestors in the distant past without supporting evidence from historical documents.8
In other words, “To answer a specific question about individual ancestry, you need to supplement your … genetic information with reliable historical records.”9
No argument there.
SOURCES
- Ashik Siddique, “DNA Ancestry Tests Are ‘Meaningless’ for Your Historical Genealogy Search,” Medical Daily, posted 7 Mar 2013 (http://www.medicaldaily.com/ : accessed 13 Dec 2014). ↩
- David Balding, Mark Thomas and Tabitha Innocent, “Sense About Genetic Ancestry Testing,” Sense About Science (http://www.senseaboutscience.org/ : accessed 13 Dec 2014). ↩
- Ibid. ↩
- Ibid. ↩
- See e.g. Judy G. Russell, “Admixture: not soup yet,” The Legal Genealogist, posted 18 May 2014 (https://www.legalgenealogist.com/blog : accessed 13 Dec 2014). Also, “Playing with percentages,” posted 24 Nov 2013; “Those pesky percentages,” posted 27 Oct 2013; “DNA disappointment,” posted 15 Sep 2013. ↩
- Balding, Thomas and Innocent, “Sense About Genetic Ancestry Testing.” ↩
- Ibid. ↩
- Siddique, “DNA Ancestry Tests Are ‘Meaningless’ for Your Historical Genealogy Search.” ↩
- Ibid. ↩
It’s interesting that people would challenge genetic genealogy as analagous to astrology when in fact there are respectable companies trying to convince us that copying from other peoples’ trees (whether we do it ourselves or use the company’s magic search button) is the way to go.
It seems to me that if there is anything we might want to challenge, it’s that.
I don’t mind copying trees, Israel — as long as the tree information is accurate on both side of the copying! In other words, combining all of the evidence (and not just wishful thinking!).
It is a shame that so many newspapers misrepresented the original Sense About Science report and gave the false impression that genetic genealogy is akin to astrology when in fact it was the misleading inferences made about our deep ancestry that were being criticised and were quite rightly being compared to astrology.
I was asked to write a follow up blog post for Sense About Science to clarify the position about genealogical DNA testing which your readers might find of interest:
http://www.senseaboutscience.org/blog.php/41/sense-about-genealogical-dna-testing
This blog post was reviewed prior to publication by the scientists who wrote the original report and also by many members of the genetic genealogy community.
I also wrote a blog post when the report first came out which provides some background about the misleading newspaper stories that prompted the scientists to write the report in the first place:
http://cruwys.blogspot.co.uk/2013/03/sense-about-genetic-ancestry-testing.html
Thanks so much for the additional information, Debbie!
Thank you Judy Russell and Debbie Kennett for the information and answers you have provided. Hope this helps a little for those who have heard so much misinformation regarding DNA testing in the genealogy world. For the people I have encouraged to test the responses are “This is something the government wants” down to “What if I find my Dad is not really my Father?”. The first is paranoia and the second makes me wish I had kept my mouth shut!
It always makes me laugh when people say the government wants our DNA. Trust me — as a former prosecutor I can assure you — if the government wants your DNA, it will get it, and in a way that doesn’t involve a genetic genealogy company (since it can’t use that DNA in court anyway).
Judy, your statement: “… the percentages reported in the ethnicity estimates by DNA testing companies are just that: estimates, and nothing more.5 They are, to be blunt, cocktail party conversation starters and not a reason to do DNA testing at all” is not quite right either, I believe.
If you are making a statement about empiricism then yes, all genetic tests are estimates.
One could thus say the same thing about health or medical tests.
When one compares your genotype against a pool of other people it is possible with high confidence to tell how alike you are to subgroup A compared for instance to subgroup B.
The problem with ethnicity estimates lay with the consumer more so the company, at least if by company we limit ourselves to the more reputable outfits like 23andMe or AncestryDNA. (Note that I believe BISDNA, a cause of much consternation in the UK where SenseAboutScience is based, is not as circumspect regarding “ethnicity” as the two companies I recommend.)
People almost always start this process (of taking a DNA test and looking at the results) with an overly simplistic understanding of themselves and their past.
I suspect our cultural aversion to teaching about human evolution has part to do with this. Also the racial/caste divides inch themselves into these discussions.
Genetics is not a simple discipline and the topic presents a very steep learning curve for the layman interested. I’ve come across plenty of egregious articles in the media that are as misleading or even more so than the advertising material from DTC genetics testing companies.
Sure it’s possible to tell whether I’m like subgroup A as opposed to subgroup B. What isn’t possible — not now and perhaps never — is being sure you really know what subgroup A represents as opposed to subgroup B. Every reference population used for the ethnicity estimates beyond the continental level is a guess. Some may be better guesses than others, but it’s still a guess. You’re comparing living people to living people, and that means of necessity that you’re guessing that this specific group of living people is like the population of, say, Denmark 1000 years ago.
I have taken an autosomal DNA test and am now considering a MtDNA test but a bit confused by the options. A simple (cheap) HVR1 and HVR2 test will only be $69 but a full sequence will be $139 (on sale). The brief list with the tests states that the first will not be useful for genealogy only the more expensive test will. I can afford the less expensive one so it will be that or nothing at present, but I wonder if this will be a waste of my time and money.
In general, an mtDNA test is advisable when there’s a specific question you have that can be solved with that type of DNA. Examples would be a family story about direct female-line descent from a full-blood Native American woman (the tested person should show a Native America mtDNA haplogroup) or descent from one of two wives and needing to prove which one (matching mtDNA with a known direct female-line descendant of one of the two wives). Otherwise, because mtDNA changes so little over so many generations, it’s more of a curiosity than a genealogically useful test. That being said, I love seeing people test because the more tests that get into the database, the more useful it will be for the cases where it’s needed!
I have found there is so much confusion…I have one cousin who couldn’t understand the point of testing as our ancestors hadn’t tested. It didn’t seem to dawn on her we carry some of their DNA and if we have it in common it is from them. She didn’t realize that we could also partially backwards engineer them with our inherited DNA. I bet it did seem like Astrology to her until I explained how it works.
Oh yes… and when the mainstream media does nothing but spread fear and hype it sure makes our job explaining it so much harder!
I’m struggling to learn the ins and outs of DNA testing for genealogy and know that I’m still a long ways from understanding the field. Meanwhile, my daughter is very anxious to have her DNA tested. We adopted our daughter from St. Petersburg, Russia and were told that her birth mother was ethnic Russian (although I suspect that may not be correct since her birth name was Ukranian). However, we know nothing about her birth father. Our daughter suspects that she may be partially Asian (not terribly unusual even in European Russia) although this is not readily apparent. She does have olive skin and dark hair and eyes and it seems likely that these traits may well have come from her birth father as many of the Russians from St. Petersburg have very fair skin and light hair and eyes.
Would the autosomal DNA testing give my daughter any clues as to her ethnic heritage or would it be a waste of money. Obviously, we can’t do y DNA testing as she is female and mtDNA is not going to give her any information on her father’s ancestry.
Any suggestions that might help my deserving daughter learn something about her ancestry? Thanks!
Autosomal DNA would certainly provide some hints at a broad brush level, and certainly should show whether she has Asian ancestry. It’s good on the continental level, not so good on a smaller scale (Europe versus Asia, sure; Russia versus Ukraine, maybe not). Of all the companies, the one that has the best ethnicity estimates right now is probably 23andMe, but really they’re all estimates and you can test with any company and then run the raw data through the ethnicity calculators at the free site Gedmatch for more analysis.
Judy – thanks so much for your great – and super fast – response! She will be glad to hear that the autosomal testing may show the broad brush, which is really all she is wanting to learn. I just discovered your blog a few days ago and I am loving it!
Glad I can help, and sure hope she gets the answers she wants.