About the SA Blog Network



Opinion, arguments & analyses from the editors of Scientific American
Observations HomeAboutContact

Genetic ancestry testing is an inexact science, task force says

The views expressed are those of the author and are not necessarily those of Scientific American.

Email   PrintPrint

genetic ancestry testing task force reportAs store-based distribution of genetic health tests remains stalled pending federal review, genetic ancestry tests are also drawing widespread concern from experts.

A task force report commissioned by the American Society for Human Genetics and led by Charmaine Royal of the Institute for Genomic Sciences & Policy at Duke University has called for better research and reporting in both academic and public genetic ancestry tests.

Some 40 companies currently offer personal genetic ancestry tests, but results and interpretations are troublingly variable, the task force found. "The very concept of ‘ancestry’ is subject to misunderstanding in both the general and scientific communities," the task force members wrote.

Scientists who research population histories have trouble sorting out definitive answers about geographic origins and ancient migrations—and they generally work with larger, more statistically weighty samples rather than with data from a single individual. Although databases of reference genetic sequences are growing rapidly, many of the major studies have been done from the Human Genetic Diversity Panel, with only about 1,100 genetic samples from across the globe, which means that even much of the vetted academic work on human ancestry is still far from conclusive.

For most individuals, genetic ancestry tests hold the promise of pinpointing where their ancestors came from. Unfortunately, science has not yet caught up to this expected—and sometimes promised—level of precision. The estimates of geographic ancestral origins are based on basal "ancestral" populations from, for instance, Northern Europe, Asia, Africa or North America. However, as the task force report authors pointed out, most of these genetic population maps are actually based on best estimates of original populations, because "we do not have the ability to sample ancestral populations." Instead, proxy populations have been used, including the Yoruba people of western Africa to represent most African-American ancestry. So even if an individual’s ancestors did hail from somewhere in western Africa, if their particular lineage (as captured in their sample) does not match up with reference sequences, some tests will turn up with no family tie.

The variety of testing methods—based on mtDNA, Y chromosome markers and autosomal DNA variants—and analyses can produce results with vastly different interpretations. Scans for haploid genetic markers can indicate a common paternal ancestor, but "they reflect only a fraction of any person’s total genetic ancestry," the authors noted. And even autosomal marker scans, which "represent a much greater proportion of genome history," still miss large portions of ancestral code because as ancestors become more distant, fewer of their genetic indicators remain, decreasing the odds that they will be found in a DNA test, Royal and colleagues reported.

With all of these big unknowns, even with the best statistics, "rarely can definitive conclusions about ancestry be made beyond the assessment of whether putative close relatives are or are not related," such as in a paternity test. So any link to specific historical figures uncovered by genetic ancestry testing, whether it is Queen Elizabeth I or Ghengis Khan, is "merely speculative."

Margin of error and statistical uncertainty are expected and usually accounted for in assessing results in an academic setting. But for results on an individual ancestry test, "interpretation often is key," the report authors noted, "because the information that is presented might have direct psychosocial and other implications for the individual." Individual and even political decisions (such as claiming rights to heritage or even citizenship) are increasingly based on genetic ancestry findings, but legal precedents have yet to be consistent in interpreting these results. The task force called for greater standardization in analysis methods but also in reporting, so that ancestry results would be delivered—and perhaps interpreted—with the proper amount of essential uncertainty.

As sales of genetic tests for health and medical profiles in brick-and-mortar stores remain in limbo pending review by the U.S. Food and Drug Administration, both those tests and genetic ancestry tests remain available for the public to purchase online.

The report authors were confident, however, that the field would continue to grow in both popular participation and scientific significance, making an impact on how people view ancestry, geography and race. It could eventually "dispel the notion of race in humans." Or it could just prompt more people to ask their doctors what their newfound ancestral identities mean for their chances of getting hypertension (a use of these tests the authors did not support).

The report was published May 14 in The American Journal of Human Genetics.

Image of human migration patterns and major mtDNA haplogroups (as seen from Antarctica) courtesy of Wikimedia Commons/

Rights & Permissions

Comments 10 Comments

Add Comment
  1. 1. JamesDavis 7:13 am 05/15/2010

    With the scientific knowledge we have today and all the major flaws it seem to possess, how can they still claim that we all evolved from one species or one person? If our current science is this flawed, how flawed do you think the science is going back millions of years? I think this would be considered "major speculation".

    Link to this
  2. 2. TMK 11:05 am 05/15/2010

    @JamesDavis – Comparing individual DNA sequences is difficult to reach any statistical conclusion. Because there is so much potential variation between individuals. A lot of the genetic markers these studies are testing are highly variable (which is why they are useful). But a lot of the variation is not informative. In other words, we need lots of statistics and because of the high level of variation, it can be chalked up to random chance that people share ancestry. So, in a way – you are correct – it’s a bit of a crapshoot.

    BUT (!) – that is dealing with modern day and current variation. Where matings are not regional and migration patterns aren’t totally understood.

    It is actually easier to look at lineages that are more ancient – because we can use sequence markers that are more conserved (less variable). And, often the sequences researchers look at are subject to evolutionary pressures (not so much for the highly variable regions used between modern day individuals).

    The science isn’t flawed, it is an ongoing quest to get more accurate. If people like you jump on every chance you get when science proclaims we don’t know all the answers, then our society will continue to descend into idiots and mud-slingers. Science can be wrong, a work in progress, entirely correct.

    Reporting negative or inconclusive results is just as important as fully vetted undebatable results. Unless people like you can’t grasp that…

    Link to this
  3. 3. hotblack 2:03 pm 05/15/2010

    Because James, just because you don’t know everything doesn’t mean you don’t know anything. …a premise religious peoples should be familiar with.

    It’s a little like… Toyota had trouble with their gas pedals. If that gas pedal production line of thinking is flawed, how flawed is the rest of their thinking? How do I know the thing in my driveway is even a car? It looks like a car, and drives like a car, but it could be a sculpture of a car, or what if it’s really a giant toaster that just happens to not be very good at heating up bread? I think I shall out with my marmalade and rye and wait patiently.

    When you throw the baby out with the bathwater, you are free to replace reality with whatever you want.

    Link to this
  4. 4. bobsmith1234 2:58 pm 05/15/2010

    @JamesDavis – Does science have flaws? Absolutely. But science is a progressive self correcting process. That’s what makes it so great. It’s foundation is not based on an argument from authority, but instead it relies on the truth, however appealing or unappealing it may be. Also, you should be skeptical, of everything; however, you must be willing to accept notions other than your preconceived ones.

    It is usually logical to follow the preponderance of evidence, as in the case of evolution.

    Link to this
  5. 5. drollere 4:14 pm 05/15/2010

    no behavioral scientist claims that modern humans are descended from a single ancestor — that is a biblical claim. most paleoanthropologists, relying on genetic calculations, assume that modern humans evolved from a "founder group" of a few thousand individuals; this group lived approximately 150,000 years after the estimated time of "mitochondrial eve" (see for example the wikipedia article).

    it’s a basic point of science literacy and simple critical thinking to know how far to accept evidence and how far to apply statistically based genetic findings. the article does not do an exemplary job of clarifying where contemporary genetic reports fail in that regard, and I’d guess there is variation across vendors in how much guidance the reports provide. but the article does signify (without explicitly mentioning it) the larger and more basic problem that the average adult citizen (and politician) in american society is woefully incapable of understanding and applying scientific evidence on any issue. the health care debate, the climate change debate, the evolution debate — all illustrate that tragic fact. mr. davis is merely a token example.

    Link to this
  6. 6. TMK 4:57 pm 05/15/2010

    I fully agree with you… but where do we begin to motivate those capable of grasping the complexities of science…

    with clever videos like this one

    Or with serious science like Science or Nature (hardly accessible to even the average scholar).

    Or with mockingly hilarious videos like this:

    Where do we start for science literacy? Let’s face it, schools are doing an absolutely abysmal job.

    Link to this
  7. 7. JamesDavis 7:49 am 05/16/2010

    Thank all of you commentators for setting me straight and I apologize for not going into greater detail in what I was referring to…I was not and I will never bask science. I know that is the only means we have in discovering the truth. When a scientist presents speculation, and paleontology and liked sciences are probably 90% speculation, as hard fact like a lawyer would in a court of law, and as most of SciAm’s articles are presented to us, then that scientist needs bashed to get them back in the right direction. One scientist claimed that we all evolved from one ape like bipedal female animal in Africa and he presented this speculation as fact and about all the people and other paleontologists who read his article believed every word he said, and now most of the science community is pushing the fact that we all came from a little tribe of monkey like bipedal animal or mammals from Africa. I do not have great knowledge in this area, but I know in my soul, that cannot be true because their have been too many bipedal like mammals on this planet and we modern bipedal mammals will not be the last. When you present speculation, and speculation is never a science, as a fact, then you are flawed to a dangerous point.

    How many science articles have you read and have been presented to the people as fact and then a year or more later another scientist disputes that article as untrue, and the scientist who presented that article seals his/her mouth, after it has screwed up a large number of people or communities? You usually say, "dammit to hell, how can you believe any of these people?", or you say, "Em bastards, they killed Kenny!"

    When you read multiple articles on the same subject and they all are different or of the same subject that have been disputed numerous times as flawed and that scientist’s findings are almost exactly as the findings that have been declared as flawed, are you just going to set by and not try to get that scientist back on the right path by pointing out their error or help them discover their flaw? Only a fool will believe everything they read or hear as fact that cannot be disputed or bashed.

    You commentators bashed me and said that my thinking on this article is flawed and I was wrong for bashing the author of this article when the author switches us in mid-stream and tells us that our learning is only a fraction of being right about a subject that has been presented over and over again for years as being factual and these flaws will screw up a lot of lives.

    Link to this
  8. 8. Bops 12:40 am 05/17/2010

    How do you compare something to something else when you have almost nothing to compare it to? You can’t.

    Even looking up family names…there are so many people with the same name. Some families name all the kids the same first born name. Second born another same name. It’s really foolish.

    They can’t even count generations because of that.

    Link to this
  9. 9. jgrosay 6:07 am 05/18/2010

    Today, many people are taking genetic tests just as a matter of curiosity and self-knowledge, thus generating a data base with no previous precedents. However, as in academy studies, the doubt remains on wether the population that ask for genetical tests is truly representative of the rest of their neighbours, a doubt that also arises when reading some reports in Am J Hum Genetics, that have no description of the sampling technique. A minimal requirement would be some kind of sampling similar to the polling surveys, even when sampling always alters the nature of phenomenon, in the way a Maxwell’s devil does. Salud +

    Link to this
  10. 10. jgrosay 6:32 am 05/19/2010

    Another fact is that when you introduce more deep studies and more refined techniques, you’ll be able to detect differences even in identical twins, in issues as epigenetics and gene expression and silencing

    Link to this

Add a Comment
You must sign in or register as a member to submit a comment.

More from Scientific American

Email this Article