Reinforcing Race in Law Enforcement
Criminal identification practices have evolved over the past two hundred years from improvised, vague descriptions noting a person’s birthplace or demeanor (Sankar 1992, 81–82) to automated, scientific biometric and genetic technologies capable of reliably and predictably distinguishing one individual from among millions. Facial recognition software can pick out a deadbeat dad from crowded stands at a sporting event (Woodward 2001; Trigaux 2001), and DNA typing can identify a murderer from the saliva left behind from licking closed an envelope (Gillespie 2003). Still, as impressive as this progress has been, one limit unites the new with the old. Identifying a wrongdoer—whether by birthplace or genetics—requires already knowing him or her. In all identification technologies that rely on matching unknown against known, there must already have been an encounter that generated a record against which the information produced in a subsequent encounter can be compared. The dad in the stands is identified as deadbeat because the database against which the new images are compared has already stored and labeled his image as such. The saliva on the stamp provides damning evidence because its analysis matches a DNA-typing profile police already have on file. But there is a new technology, forensic DNA phenotyping (FDP)—also sometimes referred to as ancestry profiling or phenotypic profiling (Cho and Sankar 2004; Ossorio 2006)—that promises to get past this limit and produce an identity sui generis. By any name, its potential utility to police is clear; however, its reliance on race to predict identity also makes it controversial (Newsome 2007).
Forensic DNA Phenotyping (FDP)
FDP analyzes DNA left at a crime scene to locate genotypes linked to ancestry and physical appearance, such as eye or skin color, and uses these genotypes to assign race and predict appearance. Researchers anticipate being able to add features for outwardly apparent behaviors, such as gait or a predisposition to smoking, or, according to one account, “physical features such as the space between the eyes, the shape of a jaw—all the things a sketch artist might want to know” (Spagnoli 2007). Imagine the advantages to police of being armed not only with the clues that a crime scene routinely provides, such as the crime’s location, timing, and method, but also with the knowledge, as the senior official in charge of Great Britain’s National DNA Database (Staley 2005, 31) suggested, that they are looking for “a 6 ft 3 in man with red hair and a tendency to obesity.” The chief scientific officer of a company that specializes in FDP predicted an even greater potential for the technology when he forecast, “A few years from now, we’re going to have figured out so many traits that a criminal might as well leave his driver’s license at the scene of the crime” (Sachs 2004). Traditionally eyewitnesses have been relied on to provide critical details about suspects’ appearance, but ample research has shown that they are often wrong (Cutler and Penrod 1995). Not only might FDP descriptions provide more detail than most eyewitnesses, but, restricted as it would be to analysis of DNA associated directly with the crime, FDP also would overcome the limits of eyewitnesses and not confuse a hapless (and innocent) deliveryman with the murderer.
Similar to DNA typing, FDP relies on genetic analysis, but in other important respects, the two technologies differ. In contrast to DNA typing (and to other matching technologies), FDP’s objective is to predict, rather than only confirm, the identity of a wrongdoer. FDP’s moment occurs during the investigation or arrest, when police are trying to locate the perpetrator. Once the perpetrator is in custody, DNA typing takes over. Its job is to convince the court that the defendant is the same person who left DNA at the crime scene. Also in contrast to DNA typing, FDP does not seek to demonstrate unique identity—it does not seek to determine precisely whether the perpetrator is this tall, red-haired man or that one; rather, FDP seeks to distinguish who might be the wrongdoer from who is not. Told that the perpetrator is a tall, red-haired, white male, the police can ignore all of the women and most of the men, or at least all those who appear not to be white or to have red hair.
In theory, FDP could move from this kind of group level sorting to unique identification, for example, identifying a specific tall, red-haired man. Doing so would require adding more and more features to a description until, as with fingerprints or DNA typing, the likelihood of the description matching more than one person would drop so infinitesimally low as to be judged impossible. Conducting the research that would support such an application would require enormous effort and methodological innovation. But whether FDP predicts a set of features or a unique description is less the point than that FDP predicts, rather than confirms, identity, a capability unmatched by other technologies.
Although one can see why this technology might appeal to law enforcement, it raises serious social and ethical questions. For now, at least, the primary, if not sole, feature on which its descriptions rely is race. While researchers hope to develop the capacity to predict other features, such as gait, voice, age, or facial morphology (Shriver 2007), they have yet to declare success on these fronts. With the exception of information about gender from the X and Y chromosomes, FDP analysis focuses on a set of genotypes linked to biogeographical ancestry (BGA), a term proponents have defined as “the heritable component of race” (Sachs 2004). From this “heritable component of race,” FDP purportedly predicts the perpetrator’s personal appearance. For example, in one case, analysts reported that a DNA sample revealed that its owner’s biogeographical ancestry was 85 percent sub-Saharan African and 15 percent Native American, and thus predicted that the perpetrator was an African American “of average skin tone” (Noel 2003).
FDP’s emphasis on race follows from an established interest in race in population genetics research, on which FDP is based. Thirty to forty years ago, population geneticists interested in questions such as how to estimate the rate of mutation among different human populations (Neel 1973, 1974) or how to estimate the degree of admixture among different populations (Reed 1969b, 1973) began to identify mutations that appeared to be present in one population but not in others, and they named these “private biochemical variants” (Neel 1974). Following established conventions, they described these variants, or alleles, as belonging to different races, as in the article title “Caucasian Genes in American Negroes,” which appeared in Science (Reed 1969a). Research in the intervening decades reported many more such alleles and built up an extensive body of work. In the 1990s, researchers interested in forensic applications of genetics turned to this work and incorporated its reliance on race into their own research.
Most of the work developing FDP has been done in the United States by the company DNAPrint Genomics, which developed a computer program, DNAWitness, for this purpose. (Unable to secure funding in the recent poor economy, DNAPrint ceased operating in March 2009. Plans for the company’s return to operations are currently unclear [Anonymous 2009]). But they are not alone in the field. The governments of Great Britain and the Netherlands conduct FDP analyses using their own procedures; Japan has announced plans to develop a genetic database for tracking ethnic minorities in that country (Cyranoski 2004); and the U.S. government recently funded a lab in Iowa to develop DNA tests to infer “population of origin” from forensic samples (Miller 2007). Only limited information about each program and its methods is available. DNAPrint reports using nearly two hundred genotypes to make its determinations, which probably qualifies it as conducting the most comprehensive analysis (Frudakis 2007). Nonetheless, what unites these programs is the assumption that geographic origins or ancestry can be inferred from genes and the information used to assign race and predict appearance.
Ancestry Informative Markers (AIMs)
The basis for this set of assumptions lies, in part, in science that developed from Neel’s research on “private biochemical variants,” which are roughly analogous to what are now called ancestry informative markers (AIMs) (Shriver et al. 2003). As with private biochemical variants, AIMS are also genetic markers, or alleles, whose frequencies have been shown to vary globally among human populations.
The distinctive distribution of these alleles results from the history of human population migration. As human populations grew and fanned out over the globe from Africa thirty thousand to forty thousand years ago, populations dispersed and settled farther and farther away from one another (Bamshad et al. 2004). At the same time, genetic changes occurred, such as those resulting from spontaneous mutation, that could be passed down from generation to generation. Depending on how close populations lived to one another, they were more or less likely to share these alleles. For example, they might be more likely to spread among people living on the same continent than among those from different continents. Thus the distribution of a small number of human genotypes varies geographically in a way that roughly reflects something of the history of human population migration. As a result, they can provide information about human ancestry, hence the name “ancestry informative markers.”
AIMs provide, however, only an estimate of ancestry. There is no one-to-one correspondence between the geographic region in which a population lives and AIMs that are common to it. AIMs that appear more frequently in one population also appear in others. For example, analysis of populations with ancestry in China would reveal certain AIMs that occur at a higher rate than they do among a population tracing its ancestry to Sweden, but any single member of either population might or might not have a particular AIM associated with that population, might have an AIM more frequently associated with the other population, or might exhibit none of the identified AIMs at all. AIMs are associated probabilistically with a population, not predictably with an individual. Applying population-level data to an individual, referred to as the ecological fallacy, is a common mistake in another venue of similar DNA testing conducted by commercial labs for people seeking to learn more about their families’ origins. Furthermore, not all genotypes, or alleles, that appear to be AIMs actually are. Alleles that vary across geographic regions might also have developed in relation to similar environmental exposures, for example, as have some alleles related to malaria resistance. As a result, a shared allele might reflect similar environmental exposures, rather than shared ancestry (Bolnick et al. 2007, 2008).
From AIMs to Appearance
The research relating population migration and AIMs is widely accepted. However, some uses of AIMs, such as claims about using AIMs to predict a person’s appearance, are highly controversial (Bolnick 2008). One concern is the oversimplification of the relationship between genotype and phenotype on which such claims rest. As critics stress (McCabe and McCabe 2006), the relationship between genotype and phenotype is complex. It is well understood only in exceptional cases, such as single gene diseases like Huntington’s disease or cystic fibrosis, where the presence of a specific, single, changed gene predicts the disease, virtually without exception. Historically, scientists assumed that more conditions would mirror the single gene model and that scientific advances would proceed by identifying a limited set of disease genes with treatments targeted at the associated phenotypes. But these assumptions are increasingly being proved wrong. Instead, researchers are discovering complex, highly contingent relationships between genotype and phenotype that challenge ready explanation. Some are associated with epigenetic events, which are heritable changes in phenotype or gene expression that result from influences external to changes in the underlying DNA (Riddihough and Pennisi 2001). Others remain unexplained, and in many fields, understanding of the genotype-phenotype relationship seems to recede, rather than advance, despite intensive study (Gaedigk et al. 2005; McCabe and McCabe 2006).
The fact that gene expression depends on complex interactions with the environment, broadly construed, means that the potential for similar phenotypic expression based on shared alleles is not always realized. Thus, for example, people who share AIMs, even AIMs proven to influence physical appearance, do not necessarily look the same. And, just as significant, to the extent that such a similarity is realized, insufficient research exists to provide a basis for accurate prediction of physical appearance based on genotype. Another concern about using AIMs to predict appearance lies with how race is implicated in predicting phenotype from genotype.
In FDP, specifically as formulated by DNAWitness, the move from AIMs to appearance starts with demarcating AIMs into groups that parallel “major continents of Europe, Asia, Africa and the Americas,” which proponents justify based on the claim that statistical analysis of AIMs produced this “4-population parental model” as the most parsimonious (Gabriel, Frudakis, and Thomas 2008, para. [0002]). Globally distributed AIMs thus become continental groups, and in turn, these continental groups provide the basis for assigning BGA, defined as “the heritable component or race.” Thus, through these steps, the broad diversity and subtle differences among groups—what some might consider a hallmark of relative youth of the human species—is reduced to four categories. At times, supporters of FDP argue that these categories are not the same as races (Gabriel 2005), but defining BGA as “the heritable component or race” seems to contradict this claim, as does using BGA as the link from AIMs to appearance.
In any case, BGA provides the link to appearance in the sense that analysis moves from BGA labels, such as European or African, to popular race categories, such as white or black. Some descriptions stop here, as in one provided for a suspect in a Boulder, Colorado, murder case, in which FDP predicted that the perpetrator would exhibit “features common to Hispanics or Native Americans” (Frudakis 2007, 607). Others add details about the value or shade of skin color, as in the description “light skinned black male” (Johnston 2006). For the most part, however, a person’s description is conveyed implicitly through the race label.
This approach to predicting identity is troublesome for many reasons, one of which is that it does not take into account a widely recognized feature of race, which is that the meaning of race labels varies across time and place. For example, people labeled in the 1980s in Great Britain as white were, by the mid-1990s, labeled Middle Eastern (Aspinall 1998; Forensic Science Service 2005), a shift that suggests that labels are transitory expressions of political relationships, rather than enduring biological categories. In admixed populations, which are populations that result from relatively recent mixing of previously distinct populations, such as Africans and Europeans might have been prior to the enslavement and transfer to the New World of Africans by Europeans, the problem of predicting appearance based on genotype is compounded and the chances of useful prediction even more diminished. Thus, although AIMs that influence physical appearance are well documented and their geographic distribution roughly parallels population migration mapped for the reasons outlined here, their capacity to predict appearance is limited.
Nonetheless, it is also important to grasp that this capacity is not entirely illusory and to understand what makes this so. Under the following conditions, FDP can work. To the extent that continental ancestry and popular race categories overlap, and to the extent that the person to be identified physically conforms to local racial stereotypes, which, in turn, need to overlap with established BGA groupings, an FDP description can predict a suspect’s identity. Similarly, say you want a rose but know only how to ask for a red flower. In certain instances, you will get a rose. This does not mean that all roses are red or that all red flowers are roses, only that some red flowers are roses (and vice versa), and thus, in certain settings (ones in which the only red flowers happen to be roses), the two appear equivalent.
FDP can work in an analogous way. If FDP provides a description that coincides sufficiently with the popular race categories held by the people looking for a suspect and with the population to be scrutinized for suspects, a description such as “light-skinned black man” might point toward the right person. This does not mean that it will always, usually, or, importantly, predictably do so. It could just as well be the case that a DNA sample described by analysts as belonging to someone who is Hispanic, or black, or white leads nowhere because the person in question does not appear Hispanic, or black, or white to the people who are looking for him. But the fact that it is likely to fail often does not mean it cannot sometimes succeed, and this is important to understand should debate over FDP become a public issue. In other words, a categorical claim that FDP cannot work is inaccurate. A stronger position highlights, instead, that it does not work predictably and that, as discussed later, it relies on stereotyping to apply results.
FDP Cases
The use of FDP has generated scant public information, but details about a few cases are available. The most fully described case, and the first official police use of FDP in the United States, involved a Louisiana man, Derrick Todd Lee. A series of murders in southeastern Louisiana reported over the late 1990s to early 2000s demanded police attention when, in 2002, four murders in a row occurred that seemed to be attributable to the same man. A multiagency homicide task force was formed, and scores of officers enlisted in the effort to locate the perpetrator. Despite spending over $500, 000 in overtime, however, the police had no luck arresting a suspect (Mustafa, Clayton, and Israel 2006, 111, 151).
Eyewitness accounts from women who thought they might have been attacked by the same man but who managed to escape, or by people who had been in the area when confirmed murders had occurred, led police to believe that they were looking for a white man, perhaps driving a white van. Federal Bureau of Investigation profiling confirmed that the murderer was likely to be white. These leads failed to turn up a viable suspect. Frustrated by the continuing murders and pressured by the growing expenses of the task force, police decided, in early 2003, to consult DNAPrint. Police submitted twenty-one samples to the company: one from the killer’s crime scenes and twenty additional samples from people known to the police (Mustafa, Clayton, and Israel 2006). This selection would allow them to use the accuracy of reports on known samples to judge the accuracy of the one report they cared about. The results came back ten days later. The twenty known samples were all accompanied by descriptions judged to be accurate, which implied that the twenty-first would be, as well. A spokesperson for the company reported that the ancestry of the person represented in the twenty-first sample was “85% sub-Saharan African and 15% Native American” and that the man “could be Afro-Caribbean or African American but there is no chance that this is a Caucasian” (Newsome 2007). Skeptical, police asked the analyst if he was sure, to which he replied, “I’m positive. You’re wasting your time dragneting Caucasians; your killer is African American” (Simons 2003).
Police finally accepted the description and used it to reorient their search. Less than two months later, they had identified Derrick Lee Todd as their man and taken a DNA sample from him to compare to those collected from the murder victims (Mustafa, Clayton, and Israel 2006). Todd had been known to police as a sexual predator and had been previously arrested and jailed, but for a variety of reasons, including his race, police had dismissed him as a suspect in these murders. On the basis of a DNA-typing match, Todd was arrested. Subsequently, he was tried and convicted for two of the seven murders possibly attributable to him. He is currently in jail, appealing his 2004 death sentence.
A second FDP case is more difficult to trace in full and is still not resolved, but is interesting nonetheless for its broader implications. The case is a rape and murder of a sixteen-year-old girl who had lived near Kollum, a small town north of Amsterdam, in the Netherlands. On the way to a dance in summer 1999, the teenager was dragged from her bike, raped, and her throat slashed (Reijnders 2005). The incident occurred near a hostel for asylum seekers, many of whom were from the Middle East or north Africa. Some in the town were quick to suspect a resident of the hostel, and the incident rapidly took on political overtones. A visit by a controversial Dutch television personality led to the development among townspeople of what one account described as a “lynching mentality” (Reijnders 2005, 636), and another commented that the response was in danger of becoming a “witch hunt against asylum seekers” (Fekete and Hoppe 2000). Officials at town meetings convened to discuss the issues were pelted with eggs. The event triggered what became a national movement in the Netherlands to change the government’s policy of dispersing asylum seekers throughout the country, including in less urban areas accustomed to a homogenous society populated by northern European Dutch (Fekete and Hoppe 2000).
The police pursued hundreds of leads in the case, but few turned up that linked the crime to the asylum seekers. A 2000 report stated, “Not one [lead] mentioned a dark-skinned person near the scene of the crime” (Fekete and Hoppe 2000). Four residents from the hostel were brought in for standard DNA typing, but no matches were produced. As of early 2009, no arrests have been made.
There are no detailed accounts of how circumstances moved along after the failed DNA typing, perhaps because what happened next appears to have been illegal. One article states that the attorney general of the province in which the crime occurred decided to seek information about the perpetrator’s possible ethnic identity through DNA analysis (M’Charek 2008, 400), while another suggests that the director of a forensic laboratory at a nearby university independently contacted authorities to obtain a sample for ethnic origins analysis (Hoekstra 2007, 10). In any case, the results indicated that the presumed perpetrator was of western European parentage, a finding that pointed away from the asylum-seeking population and toward a native Dutchman. Once made public, this information helped to deflect attacks against the asylum seekers. Eventually the incident moved the Dutch government to request that its justice department draft an amendment to the Dutch DNA law allowing FDP. The law, which passed in May 2003, remains the only statute of its kind in Europe.
Previously, in line with other European countries, Dutch law expressly prohibited forensic DNA analysis linked to personal characteristics and permitted only standard DNA typing (Koops and Schellekens 2006). The new law permits FDP analysis restricted to features that have been visible since birth, including gender (geslacht), race (ras), and population or community (bevolkingsgroep) (Korthals 2001–2002). The restriction to visible features is meant to limit the possibility of revealing genetic information to the suspect of which he or she is not aware such as a predisposition to some illness. The law also stipulates that such phenotyping is permitted only “if all other investigative tools have failed to lead to a suspect” (van der Beek 2004, 293).
Whether the Dutch case will offer the same proof of principle that the Lee case offered remains to be seen, when and if the perpetrator is caught. Nonetheless, it is significant that absent such proof, the case still led the government to amend DNA law. In light of the fact that the perpetrator has not been caught, this testifies strongly to a belief that one can infer racial or ethnic identity from genetics.
Conclusion
FDP helped to catch a brutal serial murderer in Louisiana and to defuse the threat of ethnic violence in the Netherlands. This sounds useful, so what downsides offset these potential benefits? They fall into two groups. First are the technical and logical limits to the reliability of FDP predictions and thus the utility of the technology. As discussed, the presence or absence of a particular allele or set of alleles in a person’s DNA in itself cannot infallibly reveal a person’s ancestry. While research has repeatedly demonstrated that certain alleles are more common in certain populations, it provides no basis on which to predict whether any particular individual in that population will or will not have a certain allele. Furthermore, advancing research in fields such as epigenetics suggests that predicting appearance (phenotype) based on genotype (AIMs) is a risky endeavor.
The second set of drawbacks concerns the practical implementation of FDP and the role of racial stereotyping in law enforcement. Considering that FDP descriptions are rather vague, for example, “light-skinned black man” or “Hispanic,” they could be used to target and detain any of a very large number of people. It seems possible that instead of making suspect searches more exact, the vagueness of FDP descriptions might make them more vulnerable to stereotyping. Of course, the same might be said of most descriptions the police are handed. Other descriptions, however, are not based on genetics. This is where the cultural legacy of DNA typing as the “truth machine” comes into play (Lynch et al. 2009). It is not hard to imagine that the illusion of certainty that surrounds the use of genetics in law enforcement, created by DNA typing, could confer legitimacy on FDP, as well. FDP proponents dismiss this problem. If the suspect’s DNA does not match the crime scene DNA that generated the description, proponents point out, the person will be released from custody (Schneider 2007). While logically this is true, it is scarcely reassuring, especially to members of racial and ethnic minorities in the United States, who, research has shown, are “approximately five times more likely than White suspects, per capita, to die at the hands of a police officer” (Correll et al. 2007, 1006). In the U.S. context, at least, it is easy to see how FDP might only reinforce bias and increase the dangers of racial discrimination.
In light of these concerns, it is important to consider further details about the forensic uses of DNA in the Netherlands. The original Dutch forensic DNA-typing law from the early 1990s had been written to fend off possible future innovations in DNA analysis that might permit ethnic profiling because such a practice seemed unacceptably reminiscent of the Nazi-era obsession with racial purity. The willingness to change the law less than a decade later likely speaks to the dramatic changes that the Netherlands experienced over that time as a result of the repeated waves of asylum seekers moving north from impoverished and war-torn states to settle in a more prosperous and generally safer country. The Netherlands, as much of Europe, is accustomed to stable, homogeneous populations and has struggled greatly with how to accommodate and integrate new settlers with substantially different beliefs and appearances seeking to put down roots (King et al. 2008; Monar 2008). Progress has been made overall in devising strategies to better integrate immigrants and asylum seekers, but the apparent interest in genetics as a way to grapple with an influx of strangers seems to be a step backward. Recent talk of eliminating the word ras (race) in the Dutch law and emphasizing, instead, geographic origins suggests an awareness of these problems. However, it is not clear how much difference this will make (Koops and Schellekens 2006). The current paradigm for FDP, originating in the United States, already claims to originate in geography, starting, as it does, with a framework based on the “major continents of Europe, Asia, Africa and the Americas” (Gabriel, Frudakis, and Thomas 2008, para. [0002]). This provenance, however, does not stop it from moving immediately to race. In other words, simply declaring a technology not to be about race does not make it so. The meaning or character of a technology emerges through its history and use. To wit, FDP is a technology firmly anchored in race—both through its connection to the population genetic research and through the racial nature of social deviance, law enforcement, and political authority in both Europe and the United States today. As such, its use will likely contribute to racialism. Arguably, this is potentially true of any technology, at least until such a future time when race has become socially, politically, and culturally meaningless. Nonetheless, this particular technology is so evidently troublesome that its rejection should be immediate and strong.
References
Anonymous. 2009. DNAPrint Genomics goes bust. GenomeWeb Daily News. http://www.genomeweb.com/dnaprint-genomics-goes-bust.
Aspinall, P. J. 1998. Describing the “white” ethnic group and its composition in medical research. Social Science and Medicine 47:1797–1808.
Bamshad, M., S. Wooding, B. Salisbury, and J. Stephens. 2004. Deconstructing the relationship between genetics and race. Nature Reviews Genetics 5:598–609.
Bolnick, Deborah A. 2008. Individual ancestry inference and the reification of race as a biological phenomenon. In Revisiting race in a genomic age, ed. B. A. Koenig, S. S.-J. Lee, and S. S. Richardson. New Brunswick, NJ: Rutgers University Press, 70–85.
Bolnick, D. A., D. Fullwiley, T. Duster, R. S. Cooper, J. H. Fujimura, J. Kahn, J. S. Kaufman, et al. 2007. Genetics: The science and business of genetic ancestry testing. Science 318:399–400.
Bolnick, D. A., D. Fullwiley, J. Marks, S. M. Reverby, J. Kahn, K. TallBear, J. Reardon, et al. 2008. The legitimacy of genetic ancestry tests: Response. Science 319:1039–1040.
Cho, M. K., and P. Sankar. 2004. Forensic genetics and ethical, legal and social implications beyond the clinic. Nature Genetics 36(Suppl):S8–S12.
Correll, J., B. Park, C. M. Judd, B. Wittenbrink, M. S. Sadler, and T. Keesee. 2007. Across the thin blue line: Police officers and racial bias in the decision to shoot. Journal of Personality and Social Psychology 92:1006–1023.
Cutler, Brian L., and Steven D. Penrod. 1995. Mistaken identification: The eyewitness, psychology and the law. New York: Cambridge University Press.
Cyranoski, D. 2004. Japan’s ethnic crime database sparks fears over human rights. Nature 427:383.
Fekete, Liz, and Mieke Hoppe. 2000. Populist anti-asylum movement born at Kollum. Independent Race and Refugee News Network. http://www.irr.org.uk/europebulletin/netherlands/asylum_seekers_refugees/2000/ak000016.html.
Forensic Science Service. 2005. Annual report and accounts 2004–2005. Birmingham, UK: Forensic Science Service.
Frudakis, Tony. 2007. Molecular photofitting: Predicting ancestry and phenotype using DNA. Burlington, MA: Academic Press.
Gabriel, Richard. 2005. Company interview: DNAPrint Genomics, Inc. The Wall Street Transcript. April 25. http://www.twst.com/ceos/ABN618.htm.
Gabriel, Richard, Tony N. Frudakis, and Matthew J. Thomas. 2008. Systems and methods for identifying and tracking individuals. World Intellectual Property Organization Patent Application WO/ 2008.005309, filed June 29, 2007, and published Jan. 10, 2008. http://www.freepatentsonline.com/WO2008005309.html.
Gaedigk, A., A. Bhathena, L. Ndjountché, R. E. Pearce, S. M. Abdel-Rahman, S. W. Alander, L. DiAnne Bradford, and J. Steven Leeder. 2005. Identification and characterization of novel sequence variations in the cytochrome P4502D6 (CYP2D6) gene in African Americans. Pharmacogenomics Journal 5:173–182.
Gillespie, Elizabeth M. 2003. New trick by cops closes old case. CBS News. http://www.cbsnews.com/stories/2003/05/30/national/main556194.shtml.
Hoekstra, Wiel. 2007. Barcode van het DNA [Of the DNA barcode]. De Academische Boekengids 61 (March): 10–11.
Johnston, Dave. 2006. The use of DNA in Operation Minstead: Metropolitan Police Authority. September 7. http://www.mpa.gov.uk/committees/x-eodb/2006/060907/10.
King, R., M. Thomson, N. Mai, and J. Y. Keles. 2008. ‘Turks’ in the UK: Problems of definition and the partial relevance of policy. Journal of Immigrant and Refugee Studies 6:423–434.
Koops, Bert-Jaap, and Maurice Schellekens. 2006. Forensic DNA phenotyping: Regulatory issues. Working Paper 002/2006, TILT Law and Technology.
Korthals, A. H. 2001–2002, Revision of the rules governing DNA analysis in criminal matters in connection with setting by observable characteristics. Presented at the Meeting of the Second Chamber, Netherlands, 2001–2002, 28 072, no. 3.
Lynch, Michael, Simon A. Cole, Ruth McNally, and Kathleen Jordan. 2009. Truth machine: The contentious history of DNA fingerprinting. Chicago: University of Chicago Press.
McCabe, Linda L., and Edward R. B. McCabe. 2006. Complexity in genetic diseases: How patients inform the science by ignoring the dogma. American Journal of Medical Genetics 140A:160–161.
M’charek, Amade. 2008. Contrasts and comparison: Three practices of forensic investigation. Comparative Sociology 7:387–412.
Miller, Raymond. 2007. Testing DNA samples for population of origin: Quarterly report. Quarter 2 Progress Report. Midwest Forensics Resource Center. http://snp.wustl.edu/snp-research/forensics/MFRC-Report-Miller-070815.pdf.
Monar, Jörg. 2008. Justice and home affairs. Journal of Common Market Studies 46(Suppl 1):109–126.
Mustafa, Susan D., Tony Clayton, and Sue Israel. 2006. I've been watching you: The south Louisiana serial killer. Bloomington, IN: AuthorHouse.
Neel, J. V. 1973. “Private” genetic variants and the frequency of mutation among South American Indians. Proceedings of the National Academy of Sciences of the United States of America 70:3311–3315.
Neel, J. V. 1974. Developments in monitoring human populations for mutation rates. Mutation Research 26:319–328.
Newsome, Melba. 2007. A new DNA test can ID a suspect’s race, but police won’t touch it. Wired Magazine, December 20. http://www.wired.com/politics/law/magazine/16-01/ps_dna.
Noel, Josh. 2003. Florida lab pointed to race: Serial killer search changed course. Baton Rouge Advocate, June 4. http://www.uic.edu/orgs/uicsymrg/uicsymrg/Florida%20Lab%20Pointed%20to%20Race.pdf.
Ossorio, Pilar N. 2006. About face: Forensic genetic testing for race and visible traits. Journal of Law, Medicine, & Ethics 34:277–292.
Reed, T. E. 1969a. Caucasian genes in American Negroes. Science 165:762–768.
Reed, T. E. 1969b. Critical tests of hypotheses for race mixture using Gm data on American Caucasians and Negroes. American Journal of Human Genetics 21:71–83.
Reed, T. E. 1973. Number of gene loci required for accurate estimation of ancestral population proportions in individual human hybrids. Nature 244:575–576.
Reijnders, Stijn. 2005. The people’s detective: True crime in Dutch folklore and popular television. Media, Culture, and Society 27:635–651.
Riddihough, Guy, and Elizabeth Pennisi. 2001. The evolution of epigenetics. Science 293:1063.
Sachs, Jessica Snyder. 2003. DNA and a new kind of racial profiling. Popular Science. http://www.popsci.com/scitech/article/2004-06/dna-and-new-kind-racial-profiling.
Sankar, Pamela. 1992. State power and record-keeping: The history of individualized surveillance in the United States, 1790–1935. PhD diss., University of Pennsylvania.
Schneider, Peter M. 2007. DNA-based prediction of physical traits: A new dimension for forensic genetics, or a first step towards violation of privacy leading to genetic discrimination? Paper presented at the 22nd Congress of the International Association for Forensic Genetics, Copenhagen.
Shriver, M. 2007. Complex physical traits. Paper presented at the 22nd Congress of the International Association for Forensic Genetics, Copenhagen.
Shriver, M. D., E. J. Parra, S. Dios, C. Bonilla, H. Norton, C. Jovel, C. Pfaff, et al. 2003. Skin pigmentation, biogeographical ancestry and admixture mapping. Human Genetics 112:387–399.
Simons, Dana Hawkins. 2003. Getting DNA to bear witness. U.S. News and World Report, June. http://www.usnews.com/usnews/culture/articles/030623/23dna.htm.
Spagnoli, Linda. 2007. Beyond CODIS: The changing face of forensic DNA analysis. Law Enforcement Technology, http://www.officer.com/print/Law-Enforcement%20Technology/Beyond-CODIS/1$37894.
Staley, Kristina. 2005. The police national DNA database: Balancing crime detection, human rights and privacy. Glossop, UK: GeneWatch UK.
Trigaux, Robert. 2001. Cameras scanned fans for criminals. St. Petersburg Times, January 31. http://www.sptimes.com/News/013101/TampaBay/Cameras_scanned_fans_.shtml.
van der Beek, C. P. 2004. Evolution of the Dutch DNA-law. Der Kriminalist. http://www.dnasporen.nl/docs/literatuur/The-Dutch-DNA-law-final-version.doc/.
Woodward, John D. 2001. Super Bowl surveillance: Facing up to biometrics. Santa Monica, CA: RAND Corporation / Arroyo Center.