www.sciencemag.orgSCIENCEVOL 31922 FEBRUARY 2008
the motivator of its generation, while over-
looking the possibility that an academic scien-
tist may falsify data in order to procure a grant,
ensure tenure, or otherwise enhance stature or
financial position. Generalizing about indi-
viduals on group membership in this way is
the intellectual equivalent of bigotry.
Bolnick et al. believe that anyone who
says they belong to a group should belong to
that group—regardless of whether or not their
deep ancestors (as reported by DNA tests)
were part of the parental population associ-
ated with that group. The irony is that we do
not disagree. In some cases, genetic testing is
simply not relevant—not because it is flawed,
but because it reports only one aspect of
“race” or “ethnicity.” Genomic ancestry tests
demonstrate that admixture is the rule rather
than the exception and hence support that idea
that human-derived notions of “race” are
based on the subjective and ever-changing
concepts of social and political identity.
Chief Scientific Officer, DNAPrint Genomics, Inc., Sarasota,
FL 34243, USA.
1. DNAPrint Genomics, Frequently Asked Questions
2. T. Frudakis, Molecular Photofitting: Predicting Ancestry
and Phenotype from DNA (Academic Press, Burlington,
3. M. Jobling, M. Hurles, C. Tyler-Smith, Human
Evolutionary Genetics: Origins, Peoples, and Disease
(Garland Publishing, New York, 2004).
4. M. Reidla et al., Am. J. Hum. Genet. 73, 1178
5. P. Underhill et al., Ann. Hum. Genet. 65, 43 (2001).
OUR PURPOSE IN WRITING THIS POLICY
Forum (19 October 2007, p. 399) was not
to persuade the public that tests for genomic
ancestry are wholly illegitimate, as Frudakis
assumes. Rather, it was to call attention to
this influential commercial enterprise and
the need for consumers (and the public) to
better understand the capabilities and limi-
tations of the available tests. We also hoped
to inspire genetics and anthropological
associations to discuss these issues with
LETTERS I BOOKS I POLICY FORUM I EDUCATION FORUM I PERSPECTIVES
The move not taken
edited by Jennifer Sills
The Legitimacy of Genetic Ancestry Tests
I FOUND THE POLICYFORUM “THE SCIENCE AND BUSINESS OF GENETIC ANCESTRYTESTING”
(D. A. Bolnick et al., 19 October 2007, p. 399) to be a shallow discussion of a very
Bolnick et al. criticize one company in particular—DNAPrint Genomics, Inc.—for pro-
moting the idea that race is rooted in one’s DNA (1). In fact, DNAPrint Genomics takes
great pains to explain to customers that both genetics and race are imperfectly correlated
with geography. DNAPrint’s Web site (1) and print publications (2) clearly explain that the
bases for human-derived notions of “race” incorporate genetics as well as geography,
religion, culture, and even socioeconomics. Bolnick et al., however, deceptively imply that
genetics does not constitute even a component of the equation.
Bolnick et al. also allege that genomic ancestry panels present a biased picture of non-
neutral mutations, which is not the case (2, 3). They then imply that genomic ancestry meth-
ods rely on imperfect—i.e., insufficiently large—databases and thus produce misleading
results. However, the onus on the database developer is not to build a perfect database, but
rather to quantify how imperfect the database is. DNAPrint Genomics has spent enormous
resources doing just that (2).
Likewise, Bolnick et al. focus on the
fact that alleles are continuously distrib-
uted, and that “companies sometimes fail to
mention that an allele could have been
inherited from a population in which it is
less common.” This statement is mislead-
ing. The continuity of allele frequencies is
the reason why, when estimating genomic
ancestry with respect to a given population
model, we use large numbers of well-char-
acterized markers appropriate for that
model, with algorithms capable of accom-
modating uncertainty. DNAPrint Genomics
has gone to great lengths to determine con-
fidence intervals and to quantify the bias
and mean square error of our estimates. If a
customer’s likelihood of ancestry from a certain parental population is very small, the cus-
tomer’s report will reflect this.
I also object to the implication that the detection of lower levels of “Native American”
ancestry in Asia and Europe is evidence that the AncestryByDNA test is illegitimate. Native
Americans are derived from southwestern Siberia and central Asia, an area of the world that
likely also contributed through population expansions and migrations to many other popu-
lations (2, 3). Numerous “Native American”Y and mtDNA haplogroups have been found in
Europe and Central Asia [figures 4-3 and 4-4 of (2) and (4, 5)]. The confusion lies in the
choice of the term “Native American.” Naming parental populations with descriptors based
on modern-day populations might lead to misinterpretations (2), but DNAPrint Genomics
makes every effort to explain this complex topic to a lay customer base (1).
I am also offended by the implication that scientists who work for companies are corrupt.
It is wrong to assume that any science coming from a company is suspect because money is
CREDIT: N. KEVITIYAGALA/SCIENCE
Published by AAAS
on February 29, 2008
22 FEBRUARY 2008VOL 319SCIENCEwww.sciencemag.org
We did not suggest, as Frudakis states,
that there is no connection between genetics
and societal interpretations of race. Racial
identity is shaped by a variety of factors,
including social relationships, life experi-
ences, and biological ancestry. Although
DNAPrint’s Web site states that race reflects
more than genetics, it still (as of 2008) leads
consumers to believe that race is inscribed in
one’s DNA. DNAPrint defines the “biogeo-
graphical ancestry” measured by their test as
“the biological or genetic component of
race” (1), and their underlying model rein-
forces the archaic racial view that four dis-
crete “parental” populations existed in the
past. The assertion that there is some sort of
discrete genetic component to race is prob-
lematic, and there is no evidence that only
four isolated populations existed at any
point in the evolutionary history of our
species (2, 3). Furthermore, an extensive
derivative literature makes it clear that many
people think that the AncestryByDNA test
identifies their racial makeup (4–7).
Frudakis asserts that our Policy Forum
claimed that genomic ancestry panels are rife
with biased, non-neutral mutations. We made
no such claim. Instead, we noted that some
ancestry informative markers (AIMs) involve
loci that have undergone selection (8). On
the basis of the information provided by
DNAPrint Genomics, it is clear that some
AIMs are skin pigmentation alleles and oth-
ers are blood protein alleles involved in
malarial resistance (1,4, 9–11). It is therefore
important to consider whether these markers
measure ancestry alone, or whether they also
reflect shared environmental exposures (and
thus are not always indicative of shared
ancestry). Because the AncestryByDNA test
does not differentiate between different evo-
lutionary reasons for shared alleles, the test
results may be misleading.
Frudakis then objects to the suggestion
that genetic ancestry tests are problematic
because they may yield incomplete results
due to limited sampling. He is right that per-
fect databases will never exist, and we agree
that companies should quantify the uncer-
tainty and limitations imposed by their data-
bases. DNAPrint Genomics does calculate
the bias, error, and confidence intervals of
their estimates, but incomplete geographic
sampling creates systematic bias that is diffi-
cult to quantify statistically. The sampling of
a few, widely dispersed populations for
marker selection likely influences the test’s
results. It is also U.S.-biased because it repre-
sents a specifically American racial under-
standing of human difference. Furthermore,
most mitochondrial DNA and Y-chromosome
tests do not provide any such statistics, so
consumers are often unaware that those tests
may yield incomplete or uncertain results.
Frudakis suggests that the problem with
the AncestryByDNA test is not that it
detects shared ancestry between Native
Americans and Eurasians, but that this
ancestry is referred to as “Native American.”
We agree. Because the shared alleles pre-
date the divergence of these populations and
likely originated in Central Asia, it is mis-
leading to use them as markers of “Native
Finally, it is clear that scientists who
work for companies are not inherently more
easily corruptible than academic scientists.
That said, there are differences in how aca-
demic and commercial products are evalu-
ated. Peer review may be imperfect, but it
does require academics to convince experts
that their conclusions are supported by the
data. Commercial products are not subject to
the same system of peer review, and it can be
difficult to evaluate conclusions based on
proprietary databases. Consequently, con-
flicts of interest may lead to different out-
comes in academia and the business world.
DEBORAH A. BOLNICK,1DUANA FULLWILEY,2*
JONATHAN MARKS,3SUSAN M. REVERBY,4
JENNY REARDON,8RICHARD S. COOPER,9
TROY DUSTER,10,11JOAN H. FUJIMURA,12
JAY S. KAUFMAN,13ANN MORNING,10
ALONDRA NELSON,14PILAR OSSORIO15
1Department of Anthropology, University of Texas, Austin,
TX 78712, USA. 2Departments of Anthropology and African
and African-American Studies, Harvard University,
Cambridge, MA02138, USA. 3Department of Anthropology,
University of North Carolina, Charlotte, NC 28223, USA.
4Department of Women’s Studies, Wellesley College,
Wellesley, MA 02481, USA. 5Hamline University School of
Law, St. Paul, MN 55104, USA. 6Department of American
Indian Studies, Arizona State University, Tempe, AZ 85287,
USA. 7Departments of Environmental Science, Policy and
Management, and Rhetoric, University of California,
Berkeley, CA 94720, USA. 8Department of Sociology,
University of California, Santa Cruz, CA 95064, USA.
9Department of Preventive Medicine and Epidemiology,
Loyola University Chicago Stritch School of Medicine,
Maywood, IL 60153, USA. 10Department of Sociology, New
York University, New York, NY10012, USA. 11Department of
Sociology, University of California, Berkeley, CA 94720,
USA. 12Department of Sociology, University of Wisconsin,
Madison, WI 53706, USA. 13Department of Epidemiology,
University of North Carolina School of Public Health, Chapel
Hill, NC 27599, USA. 14Departments of Sociology and
African-American Studies, Yale University, New Haven,
CT 06520, USA. 15University of Wisconsin Law School,
Madison, WI 53706, USA.
*To whom correspondence should be addressed. E-mail:
1. DNAPrint Genomics, Frequently Asked Questions
2. K. Weiss, M. Fullerton, Evol. Anthropol. 14, 165 (2005).
3. A. Templeton, Nature 416, 45 (2002).
4. K. TallBear, dissertation, University of California at Santa
5. N. Wade, The New York Times, 1 October 2002, p. F4.
6. A. Harmon, The New York Times, 12 April 2006, p. A1.
7. N. Wade, Before the Dawn: Recovering the Lost History of
Our Ancestors (Penguin Press, New York, 2006).
8. J. Akey et al., Genome Res. 12, 1805 (2002).
9. E. J. Parra et al., Am. J. Hum. Genet. 63, 1839 (1998).
10. E. J. Parra et al., Am. J. Phys. Anthropol. 114, 18 (2001).
11. T. Frudakis et al., J. Forensic Sci. 48, 1 (2003).
Soil Erosion: A Carbon Sink
THE REPORT BY K. VAN OOST ET AL., “THE
impact of agricultural soil erosion on the
global carbon cycle” (26 October 2007,
p. 626) raises two serious concerns.
First, the eroded soil is severely depleted
of its soil organic matter (SOM) pool (1–3),
which is preferentially removed by surface
runoff because it has low density and is con-
centrated in the surface layer. Along with the
SOM pool, severely eroded soil loses its
nutrient and water retention capacity. Thus,
an eroded soil generally has lower produc-
tivity even with additional input of chemical
fertilizers and organic manures (4, 5).
Because of low productivity and reduced
return of root biomass, the rate of replace-
ment of SOM pool is low on severely eroded
soils. This is in contrast to the assumptions
made by these and other authors (6–10).
Second, the process of soil erosion by
water entails three distinct stages: (i) detach-
ment, (ii) transport/redistribution, and (iii)
deposition. The first two stages lead to
breakdown of structural aggregates and
exposure of the hitherto encapsulated SOM
within the aggregate to microbial processes
with an attendant increase in mineralization
and emission of CO2(3, 11–13). During the
third depositional stage, the rate of aerobic
decomposition may be low and reduce the
emission of CO2. However, prevalence of
anaerobic conditions accentuates methano-
genesis and denitrification leading to efflux
of CH4and N2O (14), with the relative global
warming potential, respectively, of 21 and
Letters to the Editor
Letters (~300 words) discuss material published
in Science in the previous 3 months or issues of
general interest. They can be submitted through
the Web (www.submit2science.org) or by regular
mail (1200 New York Ave., NW, Washington, DC
20005, USA). Letters are not acknowledged upon
receipt, nor are authors generally consulted before
publication. Whether published in full or in part,
letters are subject to editing for clarity and space.
Published by AAAS
on February 29, 2008
22 FEBRUARY 2008VOL 319SCIENCEwww.sciencemag.org Download full-text
310 times that of CO2. Taking all three stages
into consideration, soil erosion is a strong
source rather than sink of atmospheric CO2,
and it also exacerbates the problem of non-
point source pollution and hypoxia in coastal
zones. With severe adverse impacts on agro-
nomic production and food security (15),
especially in sub-Saharan Africa and South
Asia (16), it is important that the accelerated
soil erosion is effectively controlled and
eroded soil judiciously restored.
RATTAN LAL1 ANDDAVID PIMENTEL2
1The Ohio State University, Columbus, OH 43210, USA.
2Cornell University, Ithaca, NY14853, USA.
1. F. E. Rhoton, D. D. Tyler, Soil Sci. Soc. Am. J. 54, 223
2. E. DeJong, R. G. Kachanoski, J. Soil Sci. 63, 607 (1988).
3. H. Oskarsson, O. Arnalds, J. Gudmundsson, G.
Gudbergsson, Catena 56, 225 (2004).
4. R. Lal, Science 304, 1623 (2004).
5. R. Lal, Environ. Intl. 29, 437 (2003).
6. J. C. Ritchie, G. W. McCarthy, Soil Till. Res. 69, 49 (2003).
7. W. H. Renwick, S. V. Smith, R. O. Sleezer, R. W.
Buddemier, Science 305, 1567 (2004).
8. R. F. Stallard, Global Biogeochem. Cycles 12, 231
9. S. V. Smith, W. H. Renwick, R. W. Buddemier, C. J.
Crossland, Global Biogeochem. Cycles 15, 697 (2001).
10. K. Van Oost, G. Govers, T. A. Quine, G. Heckrath, Science
305, 1567 (2004).
11. B. Schlunz, R. R. Schneider, Int. J. Earth Sci. 88, 599
12. E. A. Romankevich, A. A. Vetrov, M. E. Vinogradov, V. I.
Vedemikov, Oceanology 40, 335 (2000).
13. P. A. Raymond, J. E. Bauer, Nature 409, 497 (2001).
14. P. Jacinthe, R. Lal, Land Degrad. Dev. 12, 329 (2001).
15. D. Pimentel, Environ. Dev. Sust. 8, 119 (2006).
16. N. Borlaug, Science 318, 359 (2007).
WE AGREE THAT ACCELERATED SOIL EROSION
should not be promoted as an effective sink
for atmospheric CO2,and we want to empha-
size the need for soil erosion control and
restoration. However, our results do not sup-
port Lal and Pimentel’s assertion that agri-
cultural soil erosion is a strong source of
First, we demonstrated that the carbon
content of eroding soils is subject to contin-
ual decline as carbon is exported from the
site of erosion. It is true that selective ero-
sion of the fine fraction may further acceler-
ate this export. However, severe soil erosion
will almost always take place as tillage, rill,
and/or gully erosion, which is nonselective
(1). This local decline does not imply that
erosion is a source of atmospheric CO2
because a fraction of the eroded C was
replaced by enhanced stabilization of C in
the newly exposed soil, resulting in a net
uptake of atmospheric CO2. Although it is
true that erosion may reduce crop yields and
hence inputs of carbon at sites of erosion,
our mass-balance method accounts for all
mechanisms occurring (including a yield
decrease) and is not based on any assump-
tions about productivity or soil organic mat-
ter (SOM) replacement rate. Furthermore,
the replacement rate of SOM was low on
eroded sites, resulting in a small rather than
a large sink, but a sink nonetheless.
Second, Lal and Pimentel correctly
stated that the fate of SOM during the
various phases of the erosion-deposition
process should be considered, which is what
we did in our analysis. However, the affir-
mation that detachment and transport
of SOM inevitably lead to its accelerated
decomposition is incorrect. In contrast,
consistent with other studies, our results
showed that very little net CO2release
occurred after transport and burial within
our microcatchments (2–4).
Finally, we agree that CH4and N2O from
depositional sites warrant further research,
but our report focused on the carbon cycle
and therefore did not consider N2O fluxes.
The majority of agricultural soils are min-
eral soils, which are a sink rather than a
source for CH4(5). Although depositional
sites in agricultural fields are wetter, it is
unlikely that they will often be flooded and
reach a low enough redox potential for
KRISTOF VAN OOST,1JOHAN SIX,2
GERARD GOVERS,3TIM QUINE,4S. DE GRYZE2
1Université Catholique de Louvain, 1348 Louvain-la-Neuve,
Belgium. 2University of California, Davis, CA 95616, USA.
3Katholieke Universiteit Leuven, 3001 Heverlee, Belgium.
4University of Exeter, EX4 4RJ, UK.
1. P. J. Whiting, E. C. Bonniwell, G. Matisoff, Geology 29,
2. K. Van Oost et al., Global Biogeochem. Cycles 19,
3. S. V. Smith, W. H. Renwick, R. W. Buddenmeier, C. J.
Crossland, Global Biogeochem. Cycles 15, 697 (2001).
4. G. W. McCarty, J. C. Ritchie, Environ. Pollut. 116, 423
5. C. V. Cole et al., Nutr. Cycl. Agroecosys. 49, 221
6. D. M. Sylvia, J. J. Fuhrmann, P. G. Hartel, D. A. Zuberer,
Principles and Applications of Soil Microbiology (Pearson
Education Inc., Upper Saddle River, NJ, ed. 2, 2005).
CORRECTIONS AND CLARIFICATIONS
News Focus: “Scientists hope to adjust the president’s vision
for space” by A. Lawler (1 February, p. 564). Molly Macauley
was expressing views offered by participants at last fall’s
National Research Council forum in Irvine, California, rather
than her own opinion.
Random Samples: “When worlds collide” (7 December
2007, p. 1531). The item incorrectly stated that astronomers
estimate 1 in 1000 stellar systems host planet-forming dust
belts. In fact, that is the number of stars around which
astronomers currently see such short-lived belts; many more
stars must have them at some time in their history.
Education Forum: “Engineering education research aids
instruction” by N. L. Fortenberry et al. (31 August 2007,
p. 1175). The same data represented by the original figure
are regraphed with the y axis extended to zero (right).
TECHNICAL COMMENT ABSTRACTS
COMMENT ON “International Conservation Policy Delivers Benefits for Birds
Rolando Rodríguez-Muñoz, Alfredo F. Ojanguren, Tom Tregenza
Donald et al. (Reports, 10 August 2007, p. 810) assessed the impact of the European Union’s Birds Directive, a
conservation policy enacted in 1979, and reported evidence for positive population changes in targeted species. We
argue that their conclusions are overstatements based on unsuitable data and inappropriate analyses.
Full text at www.sciencemag.org/cgi/content/full/319/5866/1042b
RESPONSE TO COMMENT ON “International Conservation Policy Delivers Benefits
for Birds in Europe”
Paul F. Donald, Fiona J. Sanderson, Ian J. Burfield, Stijn M. Bierman, Richard D. Gregory,
Rodríguez-Muñoz et al. raise concerns about our study on the effectiveness of the European Union’s Birds Directive,
based on the provenance of the trend data used and on the predictions that were tested. Here, we show that our
results are robust to different assumptions of uncertainty surrounding trend estimates and that criticisms of the
methods stem largely from misunderstandings of the original paper.
Full text at www.sciencemag.org/cgi/content/full/319/5866/1042c
Takers, n = 2128
Non-Takers, n = 2942
Published by AAAS
on February 29, 2008