You are probably all familiar with the widely ava

www.sciencemag.org SCIENCE VOL 318 19 OCTOBER 2007 399

C R

E D

IT : N

. K

E V

IT IY

A G

A L A

/ S

C IE

N C

E POLICYFORUM

A t least two dozen companies now market “genetic ancestry tests” to help consumers reconstruct their

family histories and determine the geo- graphic origins of their ancestors. More than 460,000 people have purchased these tests over the past 6 years (1), and public interest is still skyrocketing (1–4). Some scientists support this enterprise because it makes genetics acces- sible and relevant; oth- ers view it with indiffer- ence, seeing the tests as merely “recreational.” However, both scientists and consumers should approach genetic ances- try testing with caution because (i) the tests can have a profound impact on individuals and com- munities, (ii) the assum- ptions and limitations of these tests make them less informative than many realize, and (iii) commercializa- tion has led to misleading practices that rein- force misconceptions.

The Impact of “Recreational Genetics” Although genetic ancestry testing is often described as “recreational genetics,” many consumers do not take these tests lightly. Each test costs $100 to $900, and con- sumers often have deep personal reasons for purchasing these products. Many indi-

viduals hope to identify biological relatives, to vali- date genealogical records, and to f ill in gaps in family histories. Others are searching for a connection to specif ic groups or places in Eurasia and Africa. This search for a “homeland” is particularly poignant for many African- Americans, who hope to recapture a history stolen by slavery. Others seek a more nuanced picture of their genetic back- grounds than the black-and-white dichotomy that dominates U.S. racial thinking.

Genetic ancestry testing also has serious consequences. Test-takers may reshape their personal identities, and they may suffer emo- tional distress if test results are unexpected or undesired (5). Test-takers may also change how they report their race or ethnicity on gov- ernmental forms, college or job applications, and medical questionnaires (6). This could make it more difficult to track the social expe- riences and effects of race and racism (6). Genetic ancestry testing also affects broader communities: Tests have led African-Ameri- cans to visit and financially support specific

African communities. Other Americans have taken the tests in hope of obtaining Native American tribal affiliation (and benefits like financial support, housing, education, health care, and affirmation of identity) or to chal- lenge tribal membership decisions (7).

Limitations It is important to understand what these tests can and cannot determine. Most tests fall into two categories. Mitochondrial DNA (mtDNA)

tests sequence the hypervariable region of the maternally inherited mitochondrial genome. Y-chromosome tests analyze short tandem repeats and/or single nucleo-

tide polymorphisms (SNPs) in the paternally inherited Y chromo- some. In both cases, the test-taker’s haplotype (set of linked alleles) is determined and compared with hap- lotypes from other sampled individu- als. These comparisons can identify related individuals who share a com- mon maternal or paternal ancestor, as well as locations where the test- taker’s haplotype is found today. However, each test examines less

than 1% of the test-taker’s DNA and sheds light on only one ancestor each generation (8). A third type of test (DNAPrint’s Ancestry- ByDNA test) attempts to provide a better measure of overall ancestry by using 175 autosomal markers (inherited from both parents) to estimate an individual’s “bio- geographical ancestry.”

Although companies acknowledge that mtDNA and Y-chromosome tests provide no information about most of a test-taker’s ances- tors, more important limitations to all three types of genetic ancestry tests are often less obvious. For example, genetic ancestry testing can identify some of the groups and locations around the world where a test-taker’s haplo- type or autosomal markers are found, but it is unlikely to identify all of them. Such infer- ences depend on the samples in a company’s database, and even databases with 10,000 to 20,000 samples may fail to capture the full array of human genetic diversity in a particu- lar population or region.

Commercially available tests of genetic

ancestry have significant scientific limitations,

but are serious matters for many test-takers.

The Science and Business of Genetic Ancestry Testing Deborah A. Bolnick,1* Duana Fullwiley,2 Troy Duster,3,4 Richard S. Cooper,5 Joan H. Fujimura,6

Jonathan Kahn,7 Jay S. Kaufman,8 Jonathan Marks,9 Ann Morning,3 Alondra Nelson,10 Pilar

Ossorio,11 Jenny Reardon,12 Susan M. Reverby,13 Kimberly TallBear14,15

G E N E T I C S

1Department of Anthropology, University of Texas, Austin, TX 78712, USA. 2Departments of Anthropology and African and African-American Studies, Harvard University, Cambridge, MA; 3Department of Sociology, New York University, New York, NY; 4Department of Sociology, University of California, Berkeley, CA; 5Department of Preventive Medicine and Epidemiology, Loyola University Chicago Stritch School of Medicine, Maywood, IL; 6Department of Sociology, University of Wisconsin, Madison, WI; 7Hamline University School of Law, St. Paul, MN; 8Department of Epidemiology, University of North Carolina School of Public Health, Chapel Hill, NC; 9Department of Anthropology, University of North Carolina, Charlotte, NC; 10Departments of Sociology and African American Studies, Yale University, New Haven, CT; 11University of Wisconsin Law School, Madison, WI; 12Department of Sociology, University of California, Santa Cruz, CA; 13Department of Women’s Studies, Wellesley College, Wellesley, MA; 14Department of American Indian Studies, Arizona State University, Tempe, AZ; 15Department of Environmental Science, Policy and Management, and Rhetoric, University of California, Berkeley, CA; USA.

*Author for correspondence. E-mail: [email protected] mail.utexas.edu

Published by AAAS

19 OCTOBER 2007 VOL 318 SCIENCE www.sciencemag.org400

POLICYFORUM

Another problem is that questionable sci-

entific assumptions are sometimes made

when companies report results of a genetic

ancestry test. For instance, when an allele or

haplotype is most common in one popula-

tion, companies often assume it to be diag-

nostic of that population. This can be prob-

lematic because high genetic diversity exists

within populations and gene flow occurs

between populations. Very few alleles are

therefore diagnostic of membership in a spe-

cific population (9), but companies some-

times fail to mention that an allele could have

been inherited from a population in which it

is less common. Consequently, many con-

sumers do not realize that the tests are proba-

bilistic and can reach incorrect conclusions.

Consumers often purchase these tests to

learn about their race or ethnicity, but there is

no clear-cut connection between an individ-

ual’s DNA and his or her racial or ethnic affil-

iation. Worldwide patterns of human genetic

diversity are weakly correlated with racial and

ethnic categories because both are partially

correlated with geography (9). Current under-

standings of race and ethnicity reflect more

than genetic relatedness, though, having been

defined in particular sociohistorical contexts

(i.e., European and American colonialism). In

addition, social relationships and life experi-

ences have been as important as biological

ancestry in shaping individual identity and

group membership.

Many genetic ancestry tests also claim to

tell consumers where their ancestral lineage

originated and the social group to which their

ancestors belonged. However, present-day

patterns of residence are rarely identical to

what existed in the past, and social groups

have changed over time, in name and compo-

sition (10). Databases of present-day samples

may therefore provide false leads.

Finally, even though there is little evidence

that four biologically discrete groups of

humans ever existed (9), the AncestryByDNA

test creates the appearance of genetically dis-

tinct populations by relying on “ancestry

informative markers” (AIMs). AIMs are SNPs

or other markers that show relatively large (30

to 50%) frequency differences between popula-

tion samples. The AncestryByDNA test exam-

ines AIMs selected to differentiate between

four “parental” populations (Africans,

Europeans, East Asians, and Native Ameri-

cans). However, these AIMs are not found in all

peoples who would be classed together as a

given “parental” population. The AIMs that

characterize “Africans,” for example, were cho-

sen on the basis of a sample of West Africans.

Dark-skinned East Africans might be omitted

from the AIMs reference panel of “Africans”

because they exhibit different gene variants

(11–13). Furthermore, some of the most

“informative” AIMs involve loci that have

undergone strong selection (14), which makes

it unclear whether these markers indicate

shared ancestry or parallel selective pressures

(such as similar environmental exposures in

different geographic regions) or both.

The problems described here are likely

responsible for the most paradoxical results of

this test. For instance, the AncestryByDNA

test suggests that most people from the

Middle East, India, and the Mediterranean

region of Europe have Native American

ancestry (15). Because no archaeological,

genetic, or historical evidence supports this

suggestion, the test probably considers some

markers to be diagnostic of Native American

ancestry when, in fact, they are not.

Thus, these tests should not be seen as deter-

mining the race or ethnicity of a test-taker. They

cannot pinpoint the place of origin or social

affiliation of even one ancestor with exact cer-

tainty. Although wider sampling and techno-

logical advancements may help (16), many of

the tests’ problems will remain.

Effects of Commercialization

Although it is important for consumers to

understand the limitations of genetic ancestry

testing and the complex relation between

DNA, race, and identity, these complexities

are not always made clear. Web sites of many

companies state that race is not genetically

determined, but the tests nevertheless pro-

mote the popular understanding that race is

rooted in one’s DNA (17)—rather than being

an artifact of sampling strategies, contrasting

geographical extremes, and the imposition of

qualitative boundaries on human variation.

Because race has such profound social, polit-

ical, and economic consequences, we should

be wary of allowing the concept to be rede-

fined in a way that obscures its historical

roots and disconnects it from its cultural and

socioeconomic context.

It is unlikely that companies (and the asso-

ciated scientists) deliberately choose to

mislead consumers or misrepresent science.

However, market pressures can lead to con-

flicts of interest, and data may be interpreted

differently when financial incentives exist.

For scientists, these incentives include paid

consultancies, patent rights, licensing agree-

ments, stock options, direct stock grants, cor-

porate board memberships, scientific advi-

sory board memberships, media attention,

lecture fees, and/or research support. Because

scientific pronouncements carry immense

weight in our society, claims must be carefully

evaluated when scientists have a financial

stake in them. Unfortunately, peer-review is

difficult here, because most companies main-

tain proprietary databases.

As consumers realize that they have been

sold a family history that may not be accurate,

public attitudes toward genetic research could

change. Support for molecular and anthropo-

logical genetics might decrease, and historically

disadvantaged communities might increase

their distrust of the scientific establishment

(18). These tests may also come up in medical

settings: Many consumers are aware of the

well-publicized association between ancestry

and disease, and patients may ask doctors to

take their ancestry tests into consideration when

making medical decisions. Doctors should be

cautious when considering such results (19).

We must weigh the risks and benefits of

genetic ancestry testing, and as we do so, the

scientific community must break its silence

and make clear the limitations and potential

dangers. Just as the American Society of

Human Genetics recently published a series of

recommendations regarding direct-to-con-

sumer genetic tests that make health-related

claims (20), we encourage ASHG and other

professional genetic and anthropological asso-

ciations to develop policy statements regarding

genetic ancestry testing.

References and Notes

1. H. Wolinsky, EMBO Rep. 7, 1072 (2006). 2. J. Simons, Fortune 155, 39 (2007). 3. Thirteen/WNET New York, African American Lives,

“Episode 2: The Promise of Freedom,” press release (27 July 2007).

4. P. Harris, Observer [London], 15 July 2007, p. 22. 5. Motherland, “A Genetic Journey” (Takeaway Media

Productions, London, 2003). 6. A. Harmon, New York Times, 12 April 2006, p. A1. 7. B. Hoerner, Wired 13 (2005). 8. A. Yang, Chance 20, 32-39 (2007). 9. K. Weiss, M. Fullerton, Evol. Anthropol. 14, 165

(2005). 10. C. Rotimi, Dev. World Bioethics 3, 151–158 (2003). 11. S. Tishkoff et al., Nature Genet. 39, 31–40 (2006). 12. A. Mourant, A. Kopec, K. Domaniewska-Sobczak, The

Distribution of the Human Blood Groups and Other

Polymorphisms (Oxford Univ. Press, London, 1976). 13. M. Hamblin, A. Di Rienzo, Am. J. Hum. Genet. 66,

1669–1679 (2002). 14. J. Akey et al., Genome Biol. 12, 1805–1814 (2002). 15. www.ancestrybydna.com/welcome/productsandservices/

ancestrybydna/ethnicities. 16. M. Shriver, R. Kittles, Nature Rev. Genet. 5, 611 (2004). 17. DNAPrint, Frequently asked questions, no. 1,

www.ancestrybydna.com/welcome/faq/#q1. 18. J. Reardon, Race to the Finish: Identity and Governance

in an Age of Genomics (Princeton Univ. Press, Princeton, NJ, 2004).

19. In contexts such as gene mapping and genomewide asso- ciations, genetic ancestry information can protect against confounding by population stratification or provide evi- dence of the population origin of specific susceptibility alleles (21). These applications are much narrower than determination of individual ancestry.

20. K. Hudson et al., Am. J. Hum. Genet. 81, 635 (2007). 21. M. Enoch et al., J. Psychopharmacol. 20, 19 (2006).

10.1126/science.1150098

Published by AAAS

Copyright of Science is the property of American Association for the Advancement of Science and its content

may not be copied without the publisher's express written permission except for the print or download

capabilities of the retrieval software used for access. This content is intended solely for the use of the individual

user. Copyright of Science is the property of American Association for the Advancement of Science and its

content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's

express written permission. However, users may print, download, or email articles for individual use.