By Troy Duster

Direct consumer use of DNA  tests for ancestry tracing has taken off in the last five years,   and we are not just talking about probes for first-generation genetic lineage as in the "Who's your daddy?" tests popularized on daytime television.  Since 2002, nearly a half-million people have purchased tests from at least two dozen companies marketing direct-to-consumer kits.1  The motives for testing range from the desire for ancestral links to those who lived on other continents five-hundred plus years ago to a more modest interest in reconstructing family histories.  For many African-Americans, the quest to find a link to regions and peoples of sub-Saharan Africa can take on a spiritual or even messianic quest, at least partially explained by the fact that the Middle Passage across the Atlantic during the slave trade explicitly and purposefully obliterated linguistic, cultural, religious, political and kinship ties.  The 2006 PBS television series, African American Lives, brought this quest into sharp relief.  First celebrity and later ordinary Blacks were mesmerized by stories of DNA matches that claimed to reveal or refute specific ancestral links to Africa, to Native American heritage, and surprising to some, East Asian or European populations. 

In sharp contrast, CBS' 60 minutes aired a dramatic segment in the fall of 2007 (October 7) that portrayed a direct and sharp challenge to the claims-making about such ancestry testing.  The segment began with Vy Higgensen, an African-American woman from New York's Harlem triumphantly affirming her connection to "new kin" (one of whom was a white male cattle rancher from Missouri).  But as the program unfolds, we see a disturbing cloud of doubt drift over the last part of the segment that ends with a less than subtle hint at specious claims.  A first test from the company African Ancestry, claims that Higgensen is linked to ancestors in the Sierra Leone, the Mende people.  She rejoices. "I am thrilled!  It puts a name, a place, a location, a people!"  But then she is shown the results of a second test, from another company, Relative Genetics, which claims that she instead has a genetic match to the Wobe tribe of the Ivory Coast.  She seems unruffled.  Yet a third test, from Trace Genetics, claims that her ancestors are from Senegal, the Mendenka.  Now she seems agitated, visibly concerned, confused - and most certainly disappointed that what began as a definitive match to a particular group or region of Africa has now turned into a "you pick which one you want to believe" game.

Racial classifications in 18th century Mexico.
Source: García Saiz, María Concepción. Las castas
mexicanas : un genero pictorico americano
. Olivetti

Here we have the first sally into a combined definitional and epistemological conundrum beginning with the meaning of "ancestry." While this term is typically used to refer to geographic areas where one's biological ancestors lived, with just a few minutes of reflection we can see an enormous problem to which even common sense will alert us: Which ancestors? Easy enough if we are only dealing with mom and dad, or four grandparents. We can even handle three generations back with eight great-grandparents. But if we go back six generations, that means we all have 64 direct biological ancestors. Since each of these 64 could be said to have made an equal biological contribution to our makeup, why would we choose to represent any one or two as our "real" biological lineage?  (Eight generations gives us 256 such ancestors, and twenty generations places the figure at 1,048,576.)

What can DNA tell us about our genetic lineage, and where does it fall short? What explains Vy Higgensen's multiple results from different testing sites? Flawed methodology? Partial truths hyped as definitive findings? Did the testing companies use different methods or deploy different reference populations - or both?   

Let's begin with what DNA testing can tell us about biological ancestry. There are two different tests, one for males and another for females, and each can provide relatively definitive results along one particular line of our genetic ancestry. 

Males inherit the Y-Chromosome from their biological father. The markers are sufficiently distinctive so that the test can not only identify the father, but also the father's father, and if the data were available, the father's father's father. This path to ancestry identification can go on for as many generations as data are available - which is how Thomas Jefferson (or one of his brothers) was linked to Sally Hemmings' offspring.

The test for female ancestry has an interesting parallel. We can definitively answer "Who's your mommy?" All of a mother's children inherit her mitochondrial DNA (mtDNA).  Located within the cell but outside the nucleus, mtDNA serve as the cell's energy producers but only the daughters pass it on. Thus, for a female, it is possible to trace and identify her mother, her mother's mother, etc., (along the same line as just noted for males using Y-chromosome analysis). This was the way that granddaughters were linked to their grandmothers in the aftermath of Argentina's "Dirty War" (1976-83). Thousands of young fathers and mothers "disappeared" by acts of the ruling junta, and their orphaned children were given to couples who wished to adopt.2 It was through mitochondrial DNA testing that the grandmothers were reunited with the children of their murdered or missing daughters. 

These two tales reveal not only the power of DNA ancestry testing, but their significant and consequential social and political uses as well. However, it is also vital to re-state the limitations. These two tests can identify, for example, only two of the 64 great great great great grandparents.  Indeed, only two of the next generation further back of 128 can be so identified, only two of 256, and so on.  Yet each of the other 62 or 126 or 254 ancestors contributed equally to our genetic makeup as the two we can trace by the sex-linked paternal or maternal lines.

To supplement the limitation of Y-Chromosome and Mitochondrial DNA testing, a group of researchers has come up with a procedure to discern the frequency of certain markers that are hypothesized as belonging, selectively, to our ancestors.  However, there are several blind assumptions that have to be accepted in order to have confidence in the links to ancestral populations so defined.

Unlike Y DNA or mtDNA tests, this technology examines a group's relative share of genetic markers found on the autosomes, the non-gender chromosomes inherited from both parents.

Since Ancestry-Informative Markers (AIMs) are overwhelmingly shared across all human groups, it is not their absolute presence or absence but their rate of incidence that is usually being analyzed. This is especially true when it comes to claims about continental populations. How did these markers come to represent ancestral populations of Africa, Europe, and Native America? The vast majority of these markers are not 'population specific,' as the inventor of Ancestry Informative Markers originally claimed.3  Because the companies marketing ancestry tests hold proprietary interests in their techniques, most do not make them available for possible scientific replication, and their modeling constructs are therefore undisclosed.  Thus, we are left to speculate about the threshold level of frequency that is used to determine the grounds for inclusion or exclusion, as well as what counts as a "pure" reference population.

There is a yet more ominous and troubling element of the reliance upon DNA analysis to determine who we are in terms of lineage, identity, and identification. The very technology that tells us what proportion of our ancestry can be linked, proportionately, to sub-Saharan Africa (ancestry-informative markers) is the same being offered to police stations around the country to "predict" or "estimate" whether the DNA left at a crime scene belongs to a white or black person. This "ethnic estimation" using DNA relies on a social definition of the phenotype (the observable physical or biochemical characteristics of an organism, determined by both genetic makeup and environmental influences). That is, in order to say that someone is 85 percent African, we must know who is 100 percent African. Any molecular, population, or behavioral geneticist who uses the term "percent European" or "percent Native American" is obliged to disclose that the measuring point of this "purity" (100 percent) is a statistical artifact that begins not with the DNA, but with a researcher adopting the folk categories of race and ethnicity.

The work of Evett et al (1993, 1996), Lowe et al (2001) and others suggest that there are only about ten percent of sites in the DNA that are "useful" for making distinctions.  This means that at the other ninety percent of the sites, the allele (one member of a pair or series of genes that occupy a specific position on a specific chromosome) frequencies do not vary between groups such as "Afro-Caribbean people in England" and "Scottish people in England."  But it does not follow that because we cannot find a single site where allele frequency matches some phenotype that we are trying to identify (for forensic purposes, we should be reminded), that there are not several that will not be effective, for the purposes of aiding the FBI, Scotland Yard, or the criminal justice systems around the globe in highly probabilistic statements about suspects.

When representative spokespersons from the biological sciences say that "there is no such thing as race," they mean that racial categories have no discrete boundaries, that there is nothing mutually exclusive about our current (or past) categories of "race," and that there is more genetic variation within categories of "race" than between. All this is true. However, when Scotland Yard or the New York City police force wants to narrow the list of suspects in a crime, they are not primarily concerned with tight taxonomic systems of classification with no overlapping categories. Some African-Americans have Cystic Fibrosis even though the likelihood is far greater among Americans of North European descent, and in a parallel if not symmetrical way, some American whites have Sickle Cell Anemia even though the likelihood is far greater among Americans of West African descent; but in the world of cost-effective decision-making, genetic screening for these disorders is routinely done based on common-sense versions of the phenotype.  The same is true for the quite practical matter of naming suspects.

Much like the industry of assisted reproduction in the United States, there is a complete absence of regulation or quality control of genetic ancestry testing. There is no requirement for transparency in the construction and use of reference populations. Any company can claim to provide accurate information about your ancestry. If three different companies offer three different answers (as in the 60 Minutes report noted at the outset), which one is more likely to be correct? There is no way of knowing, since we have no "gold standard" for excellence or professional self-policing. This was pointed out in Science two years ago1, and in November 2008, the American Society of Human Genetics (ASHG) issued a statement on ancestry testing that included five recommendations emphasizing the need for greater responsibility, research, explanatory clarity, collaboration and accountability by these direct-to-consumer companies.# The statement also pointedly warned of several important limitations to the scientific approaches used to infer genetic ancestry, including the false assumption that contemporary groups are reliable substitutes for ancestral

populations and the lack of transparency regarding the statistical methods that companies use to determine test results..4

While the ASHG statement calls for greater transparency, we have seen that private sector providers of ancestry testing have proprietary reasons for keeping secret their own particular combinations of key technology, software and population sampling procedures. Most are unwilling to disclose the size and composition of their reference populations. Without mechanisms to enforce transparency, there is no way of assessing the scientific basis for specific assertions of "percent ancestry." For example, until and unless there is a publicly available version of what constitutes a 10 percent European or a 100 percent African genome, claims about 80 percent ancestry cannot be fully understood or tested, much less replicated. 

Building on the ASHG recommendations for transparency, there is a need for specific policies enforced by federal agencies.  For example, the Federal Trade Commission and the Centers for Disease Prevention and Control can and should play pivotal roles in setting industry standards for what constitutes responsible and accountable practices. These agencies can promote the research necessary to identify minimal guidelines for presenting the fair uses and clear limitations of current genomic technologies. Guidelines for transparency would also include clear statements spelling out the risks associated with over-extrapolating or misinterpreting genetic ancestry results. The active involvement of regulatory agencies would provide infrastructure for the interdisciplinary dialogue necessary to create effective policies and for maintaining industry standards (Lee et al 2009).  While supporting such measures, we should not be naïve about their effectiveness, since the demands on these companies to generate profits are strong and insistent.                                           


Troy Duster, PhD, is Silver Professor of Sociology and Director of the Institute for the History of the Production of Knowledge at New York University. He also holds an appointment as Chancellor's Professor at the University of California, Berkeley.



1. Bolnick, Deborah A., Duana Fullwiley, Troy Duster, Richard S. Cooper, Joan H. Fujimura, Jonathan Kahn, Jay S. Kaufman, Jonathan Marks, Ann Morning, Alondra Nelson, Pilar Ossorio, Jenny Reardon, Susan M. Reverby, and Kimberly TallBear, "The Science and Business of Genetic Ancestry Testing," Science, 18, (19 October 2007) 399-400.

2. Penchaszadeh, Victor B. "Abduction of Children of Political Dissidents in Argentina and the Role of Human Genetics in Their Restitution, Journal of Public Health Policy, 13, 3 (Autumn), 1992), 291-305

3. Shriver, Mark D., Michael W. Smith, Li Jin, Amy Marcini, Joshua M. Akey, Ranjan Deka, and Robert E. Ferrell, "Ethnic-Affiliation Estimation by Use of Population-Specific DNA Markers," American Journal of Human Genetics, 1997, 60:957-964.

4. Lee et. al 2009

Search: GeneWatch
For centuries, human societies have divided population groups into separate races. While there is no scientific basis for this, people unquestioningly accept these classifications as fact.
View Project
The purpose of the Genetic Bill of Rights is to introduce a global dialogue on the fundamental values that have been put at risk by new applications of genetics.
View Project