13 The Difference Genomics Makes

1. Race in Public Health and the Early Human Genome Project

The formation of the Human Genome Project and its early institutionalization in laboratories across the world is a most fascinating history that has been analyzed from many different perspectives, but one of the most notable aspects of the project’s foundation was its immediate immersion in ethical, legal, and social implication (ELSI) debates. ELSI has blossomed within the US National Institutes of Health to become a powerful program reflecting on the social implications of the research priorities of its National Human Genome Research Institute. The ELSI program has made great inroads in its support of research on human variation and race. However, race was not on its agenda in the early years of the project.

The relationship between race and biomedicine was under construction throughout the late 1980s. In 1986, at the project’s launch, the National Institutes of Health deliberated over how to promote inclusion of women and minorities in science and research (Epstein 2007). Awareness of racial and gendered health disparities was growing across American social institutions, as well as in other public health systems around the world. The National Institutes of Health was part of a larger global dialogue over equity in medicine. The US Congress had devised a set of racial categories—white, Black, Asian, and Native American (Pacific Islander would later be separated out into its own racial grouping)—that were to be used in tracking disparities in racial participation in public programs, but these categories only began to be considered for use in science and medicine when the concept of health disparities took hold in public health deliberations within the National Institutes of Health and Department of Health and Human Services.4

In 1989, the Surgeon General released a report on the status of American health that designated data collection of racial health disparities a top priority of public health. Soon after, the Department of Health and Human Services (US Department of Health and Human Services 1994) cited the reduction of health disparities as one of its main goals, stating, “Minority populations are growing faster than the population as a whole. Eliminating health disparities is of critical importance in the 1990s.” In its Healthy People 2000 report, a document cowritten by the Institute of Medicine, Public Health Department, fifty-four state and local health departments, and over three hundred membership organizations, called for a new focus on minority health. Inclusion by federal racial categories was the watchword of the day (Stoto et al. 1990, 51). Policymakers moved to establish task forces on women and minority health, as well as offices for special-populations health.

This was a sea change from the one-size-fits-all way of treating human health and genetic differences that had held sway throughout the twentieth century. Groups had been seen as similar enough to stand in for any and all. Furthermore, scientists avoided working with vulnerable populations, those groups that had historically faced discrimination, and instead used “white male” DNA samples as reference points (Stoto et al. 1990, 44–45).

The Human Genome Project started out under these auspices. Funded largely by the US government, with an initial budget of $200 million per year provided by US Congress alone, the project mirrored the dominant paradigm for dealing with population variation (Burris et al. 1998).5 The project constructed a single set of human autosomes from cell lines already in existence, predominantly drawing from a set of samples the had been collected from white Americans, Europeans, and South Americans in the United States, France, and Venezuela. As chair of the subcommittee on the third Five-Year Genome Project plan, Aravinda Chakravarti, explained, initially project planners decided to “not focus on the variation, because the point is not to study variation.” They knew that sequencing was being “largely done on the European group,” but explained that this was because it was “safe and because it was a practical matter of getting the map right.” European ancestry was believed to be sufficient in constructing, as Chakravarti put it, “a human genome, not a human’s genome.”

Again, ELSI’s first Five-Year Plan did not discuss race or matters of population variation and representation. It covered issues like harms to the individual, privacy, confidentiality, employer, and insurer discrimination, and dealing with personal susceptibilities and risks (National Institutes of Health 1990). The project’s second Five-Year Plan claimed to: “(i) Continue to identify and define issues and develop policy options to address them. (ii) Develop and disseminate policy options regarding genetic testing services with potential widespread use. (iii) Foster greater acceptance of human genetic variation. (iv) Enhance and expand public and professional education that is sensitive to sociocultural and psychological issues” (Collins and Galas 1993). Though its final two goals held a promise for raising issues around race and the project’s potential to alter societal notions around it, again race and population variation were not featured. In 1992, the National Center for Human Genome Research, soon to be institutionalized as an official institute of the National Institutes of Health, issued its first retrospective, in which ELSI’s priorities were once again deemed to revolve around privacy, discrimination against individuals, and test dissemination in particular, not race, diversity, or matters affecting groups (Jordan 1992).

2. Diversity and Revitalization

As policymakers trained their efforts on finding ways to monitor and engender racial inclusion, an international team of scientists headquartered at Stanford University voiced an interest in spinning off and developing a new human genome project that would bring diversity into the picture. The Department of Energy, National Institute of General Medical Sciences, National Human Genome Research Center, and National Science Foundation sponsored a series of planning workshops. Though the project was independent of the National Institutes of Health, the Genome Project’s leadership organizations, the National Human Genome Research Center and the Human Genome Organization, publicly welcomed the project. Human Genome Organization President Walter Bodmer called the nascent project “a cultural obligation of the [human] genome project” (Roberts and Gibbons 1991).

The founders of this new project called it the “Human Genome Diversity Project.” They all saw diversity in terms of global coverage but debated how just to achieve it. They deliberated between two strategies in particular: grid-sampling the globe versus targeting preexisting groups. In the end, they decided that grid-sampling was too expensive (Roberts 1992, 1205). As lead scientist Luigi Luca Cavalli-Sforza explained it, they deigned to gather a sample of the biodiversity that would have existed in the “population defined before 1492, before the great expansions.” A longtime collaborator of Bodmer, Cavalli-Sforza said that the project was not worried about getting enough funding, but rather they needed to garner and maintain support from the US government, US policymakers, and the general public in this moment when values and priorities were drastically changing around the role of race.

Diversity Project planners criticized the Genome Project’s color-blind sampling protocols. In particular, Cavalli-Sforza and the other leaders of the Human Genome Diversity Project cited the Human Genome Project for Eurocentrism. They presented the Diversity Project as the first large-scale sequencing initiative to move beyond “Caucasoid” sequencing (Bowcock and Cavalli-Sforza 1991, 491; Roberts 1992, 1204). They held up group identity and groups self-identification, as a mode of self-determination, as the new gold standard of genome mapping, arguing that sampling populations would need to be at the forefront of sampling protocol conception (Roberts 1992, 1204).

Cavalli-Sforza told me how he embarked upon sample collection: “I made two hundred. The Chinese ones were made by a friend of mine who was working with me. Pakistan donated another two hundred, and [that donor] also was my student. Israel donated several. I didn’t want any Jews, because I didn’t want any anti-Semitic work going on, but they would have given them to me. But I took strict representatives of the Middle East, and again from France, and then all the others were from Ken Kidd from Yale.”

What project scientists didn’t realize was that the pressure to use continental classifications amenable to data collection by US federal race categories was about to render their project obsolete.

In 1993, the National Institutes of Health issued the Revitalization Act, a statute mandating the inclusion and surveillance of women and minorities in clinical research and clinical trials using the US government’s classification system (National Institutes of Health 1993).6 The agency required publicly funded investigators to tabulate enrollment by federal race domestically and internationally, and to conduct “outreach” to racial minorities. Researchers were no longer permitted to use study cost as a determinant of racial inclusion. In fact, the burden of proof was on researchers who planned not to include racial minorities to show that their study’s variables had already been shown to have the same effect across racial groups. This was the first policy to institutionalize the use of race—categorized according to the federal race classification scheme—in biomedicine nationally, via publicly funded research, and internationally, via global genome projects (Kelty et al. 2002).

The Revitalization Act had immediate effects on the Diversity Project (Reardon 2005). Project leaders reconsidered their avoidance of groups like African Americans, whom they had initially rejected for having admixed genetic ancestry.7 However, this came too late (Reardon 2005, 5). The Department of Health and Human Services, which had just begun inspecting genomic sampling protocols, ordered the National Research Council to investigate the Diversity Project’s inclusionary protocols (National Institutes of Health 1993, 1994). In 1994, Cavalli-Sforza raised the issue of racism to an audience at the United Nations Educational, Science, and Culture Organization. The Diversity Project also appointed bioethicist Henry Greeley to create a recruitment protocol designed to carefully enroll vulnerable groups.8 In order to maintain the federal support they would need to proceed, they did all they could to align their notion of diversity with the US federal government’s.

Still, the outcry from Native American groups, African Americans, and other members of vulnerable populations prevented the Diversity Project from going any further. On the international front, UNESCO’s International Bioethics Committee charged that formerly colonized people and racial minorities would be vulnerable to racist interpretations of project data (Gannett 2001).9 The National Research Council asked the Diversity Project to end international sampling altogether, instead focusing on sampling in the United States. This way the US government would be able to have better oversight on the project as well as a final say on the protocols deployed (Schull 1997). The Diversity Project’s attempts to use preexisting classifications, on the order of groups formed “before 1492,” was not explicitly racial enough in the politicized terms used by the federal government to satisfy the major players in public health and the major funders of genomics (Reardon 2005). This sent a message to the entire field to rationalize research in the specific racial terms used by the federal government. From here on, regardless of the moniker a project used, whether it be “ancestral groups,” “ethnicities,” or “major populations,” projects calibrated their groupings to the continental-racial framework mandated by the US federal policy (Panofsky and Bliss 2017, 59–87).10

3. Late Human Genome Project Developments

Just as the Human Genome Diversity Project was folding in the middle of the decade, ELSI began its investigation into the social implications of using race in genomic science. It funded two studies on race, diversity, and the representation of populations. Neither study was conducted by genome scientists, but they set the foundation for future work focused on genetic differences and race. The first study examined the effects of genetic testing on African Americans with sickle cell anemia. The second study investigated the effects of genetic testing on families at risk for sickle cell anemia, cystic fibrosis, and thalassemia—genetic diseases popularly understood to be restricted to particular races. ELSI also funded conferences in which the Genome Project and community-based organizations could come together to discuss the Genome Project’s implications for minority communities. These conferences joined scientists, social scientists and humanists, ethicists, and community leaders together to create a model recruitment protocol. They suggested that community partnership should be the foundation for project sampling.

Genome Project leaders also began their work with race. In 1996, the National Center for Human Genome Research and the Department of Energy called for greater inclusion of women as subjects in genetic research with its first-ever “Guidance on Human Subjects Issues in Large-Scale DNA Sequencing” (US Department of Energy 1996). It stated that the Genome Project would now be dedicated to “ensuring that the initial version of the complete human DNA sequence is derived from multiple donors,” “providing donors with the opportunity to make an informed decision about whether to contribute their DNA to this project,” and “taking effective steps to ensure the privacy and confidentiality of donors.” By multiple donors, these agencies were chiefly concerned about privacy protections. However, when considering how to incorporate multiple donors, diversity was stated as an overarching goal.

Soon after, the newly institutionalized National Human Genome Research Institute began embarking on its own diversity project sampling by continental groups—“the major geographic regions of the world—Europe, Africa, the Americas, and Asia”—and using the framework of US federal race categories (Collins et al. 1998, 1229–31). The Polymorphism Discovery Resource was the first large-scale sequencing project dedicated to exhibiting the world’s genomic diversity by race. Hoping to avoid racist interpretations of its data, the project pooled its racialized data into subsets of eight, twenty-two, forty-four, or ninety samples that reflected the proportional racial diversity of the total sample. The idea was to capture global diversity with the US federal categories, but to protect groups from being compared on account of their race. There was no debate about the use of these categories to cover the world’s population.

Like the Diversity Project, Discovery Resource leaders such as Francis Collins and Aravinda Chakravarti saw their project as the Genome Project’s “cultural obligation” (cf. Bodmer) finally come to fruition. However, they set themselves in stark relief to the Diversity Project by working with vulnerable populations to determine their sampling and distribution policies. Francis Collins, then director of the National Human Genome Research Institute and leader of the Genome Project and Discovery Resource, explained how discussions with Native American community leaders led to the project’s racially disambiguated policy: “At that point because we wished to include American Indians, and they felt very strongly that having samples identified as coming from Indian tribes was unacceptable, we made the decision that for that resource—which was really just there to discover variation at that point and not try to connect it to anything—that it was reasonable to have no labels” (Bliss 2012, 50).

It was of utmost importance to the Genome Project that its own attempt at mapping diversity be done at once in accordance with the federal concept of race and that concept as reflected in public understanding, specifically minority group self-determined notions.11 As project leaders stated in a report in Science: “Publicly available DNA collections contain little African material; Native American and Asian contributions are similarly scant. As a result, say NHGRI staffers, it will be essential to collect DNA from a racially structured set of donors. Once the DNA has been sampled, however, all personal and racial data will have to be removed to protect privacy—diminishing the scientific value of the project, but bolstering its ethical foundation” (Marshall 1997).

Here, it was made clear that racial inclusion and minority rights were at the head of all priorities. Indeed, the project used terms like “reflect the diversity” and “redressing Eurocentrism” to represent the project’s goals (US Department of Energy 1999).12

The private company Celera that would eventually contribute to the draft map of the human genome held a similar interest in diversification. Its leader, Craig Venter, explained that by 1998, when Celera entered the race, getting a diverse sample was essential to the project, despite pushback from his company’s board and other gatekeepers who were less aware of the change in ethical frames. “Many on the committee were very concerned that the data would be used to justify racism for the same example I gave you: take any two people; they are going to look different, and therefore the two races are different, right? But I felt it was important symbolically for the first ‘human genome’ to [be representative]. So we made decisions on a pragmatic level of DNA to cover, but out of the pool of volunteers we tried to select people to have as much sexual and ethnogeographic diversity as we could in five people.”

Venter was successful in getting racial diversity in the racial strata of, as he put it, “African-American, Chinese, Hispanic, Caucasian, etc.” Of twenty-one donors, Celera chose five representative samples along the racial lines provided by National Institutes of Health guidelines. Venter’s efforts led to more than diversification. It led to the restructuring of the very basis of global genome projects dedicated to mapping humankind, and especially their ethical and social implications for the world.

In the final years of the Genome Project, the project made its new stance on race public. In its last Five-Year Plan (Collins et al. 1998, 682–83), it stated, “additional ethical, legal, and social issues that need to be anticipated, considered, and resolved.” Finally, one of its ELSI goals pivoted on race: “Explore how socioeconomic factors and concepts of race and ethnicity influence the use, understanding, and interpretation of genetic information, the utilization of genetic services, and the development of policy” (Collins et al. 1998, 688).

ELSI also set up its own funding mechanism for research into race, ethnicity, and culture. “Concepts of Race, Ethnicity, and Culture: Examination of the Ways in Which the Discovery of DNA Polymorphisms May Interact with Current Concepts of Race, Ethnicity and Culture” called on research to ascertain: “How will individuals and groups respond to potential challenges to or validations of their racial, ethnic or cultural self-identification, based on new genetic information? What impact will the discovery of information on genetic variation have on the current categories (as defined by the government) of race and ethnicity? Will the discovery of patterns of DNA polymorphisms affect the concepts of race and ethnicity that are used by genetic, anthropological and other health and social sciences researchers?” (National Institutes of Health 1999).13

This research mechanism provided larger sums of funding than ever before for natural and social scientists, humanists and ethicists, and others focused on the study of race and society. Meanwhile, the National Human Genome Research Institute, which was in deliberation with a range of such scientists over the meaning of race, discussed creating an intramural Race and Genetics Working Group.

4. Race Under the Electron Microscope

The year 2000 was a momentous time for the Genome Project. The project moved to formally close, and to release a draft map of the human genome. The White House held a grand celebration in which project leaders Francis Collins and Craig Venter stood alongside President Bill Clinton as he interpreted the meaning of the genome project for science and the public. Clinton announced that this project’s elucidation of the human genome proved “in genetic terms, all human beings, regardless of race, are more than 99.9 percent the same.”14

His sentiment that genomics shed light on one of the world’s thorniest problems was shared by all. Venter echoed Clinton’s message stating, “I am happy today that the only race we are talking about is the human race.” He noted that race, while used in recruitment, was illegible in the mapped sequence.15 Project leader Eric Lander joined Venter and Chakravarti in telling the New York Times that race was “a bogus idea” (Angier 2000). Indeed, with the announced closing of the project, genome mappers entered the public debate over the meaning of race at full force. In interviews, Lander further stated, “The whole notion of differences between us has been about ranking people . . . We need to get that out of our heads.” Venter charged, “No serious scholar in this field now considers race to be a scientific concept” (Gee 2000).

Race, its role in genomic science as well as its general veracity, became a core phenomenon of interest for genomics, as equally displayed in its signature publication venues. On the heels of the genome project celebration, the field’s leading journal, Nature Genetics, issued an editorial calling on scientists to consider their use of race: “From now on, Nature Genetics will therefore require that authors explain why they make use of particular ethnic groups or populations, and how classification was achieved. We will ask reviewers to consider these parameters when judging the merits of a manuscript—we hope that this will raise awareness and inspire more rigorous design of genetic and epidemiological studies” (Anonymous 2000, 97–98). They deemed it every researcher’s responsibility to distinguish their definition and use of race, to note their characterization of the genome’s relationship to it, and to make plain their sampling protocols for all to see. With the Genome Project’s leadership, the journal’s editors called race a “pseudo-biological variable” that demanded careful consideration.

Following this charge, two teams of genome scientists attempted to measure race with genomic technology. Barbujani and colleagues at Max Planck Institute investigated whether racial groups akin to continental variation could be gleaned from two different sets of DNA samples. They found that continental clustering occurred in both sets, but in different ways for each. They also found that clusters were less than 70 percent predictive of individual sample origins. In other words, clear separate races were not identifiable in the genome (Romualdi et al. 2002). Meanwhile, Goldstein and colleagues at University College-London and Oxford investigated whether drug response genes could be mapped onto racial groups (Wilson et al. 2001, 265–69).16 They similarly concluded that “commonly used ethnic labels (such as Black, Caucasian and Asian) are insufficient and inaccurate descriptions of human genetic structure” (Wilson et al. 2001, 266).

Following these studies, a number of biomedical journals followed Nature Genetics in calling for contributors to define race in their methods sections, declaring suspicion about its utility as a proxy for biological variation (see, e.g., Litt 2001; Rivara and Finberg 2001). But in the emerging subfield of pharmacogenomics, it was unclear how scientists were to evaluate the need for race. The Food and Drug Administration, in 1998, had issued the Demographic Rule, a policy mandating all applicants for new drugs to detail safety and efficacy by US federal race. Like the Revitalization Act, this policy required researchers to tabulate inclusion by race. However, the policy went further in asking scientists to consider the pharmacokinetic differences in racial groups. “(1) Different subgroups of the population may respond differently to a specific drug product and (2) although the effort should be made to look for differences in effectiveness and adverse reactions among such subgroups that effort is not being made consistently.” Race was at once being declared to be different from genomic difference, yet essential to understanding drug response.

In the wake of the Demographic Rule, there was a deluge of drug response research and development by race, and such research was not limited to new drug applications but rather came from reexamining successful therapies as well (see Xie et al. 2001, 815–50). Leaders in Pharma and Biotech voiced great enthusiasm for race-based research, which would enable them to align themselves with the now dominant ethical mandate while reducing the scope of drug trials and streamlining drug pipelines (Weiss 2000).

In a dizzying set of events in 2001, drug makers published a series of articles that would set the subfield on course for race-based medicine. The African American Heart Failure Trial (AHeFT) reported a lesser response to angiotensin converting enzyme inhibitors in “black as compared with white patients” with a specific kind of heart failure (Exner et al. 2001, 1351–55). It also reported the benefits of a beta-blocker “of a similar magnitude in both black and non-black patients with heart failure” (Yancy et al. 2001). These articles showed that members of the genomics community believed that there were life-or-death differences in the biology of members of different racial groups that affected their ability to metabolize and benefit from drugs for common disease.

The outcry from scientists interpreting the genome’s meaning for race only grew stronger. However, instead of communicating a common and undisputed take, scientists made contradictory statements. In the New England Journal of Medicine, two editorials posed contrasting views (Exner et al. 2001; Yancy et al. 2001). One argued, “The Human Genome Project now gives us the power to uncover the true origins of genetic variations; linking them to race has become passé” (Schwartz 2001). The other encouraged researchers to find “genetic determinants of the racial differences, rather than [put] attention to the external phenotypic manifestations of race” (Wood 2001).17 This editorial, written by a pharmacogenomics expert who was in the process of studying racial pharmacogenomics of beta-blockers, backed the Food and Drug Administration’s belief that drug response most definitely varied by race, and that it was the researcher’s job to uncover the mechanisms behind this differential response. Despite the fact that one editorial sought to put an end to linking race to genomics, while the other encouraged elucidation of the connection, both called for greater racial inclusion and more genomic studies of health disparities.

Race was under the microscope in genomics, and especially drew attention from pharmacogenomics researchers. When the Pharmacogenomics Journal launched that year, the problem of race was front and center. Some researchers argued that “the inclusion of an ethnic minority in a clinical study may reveal information on the frequency of an unknown genetic variant which affects a drug response; such information can be of clinical importance and may increase the safety of some members of that ethnic population.” Others warned that genomic studies “should be accomplished by mechanisms based on scientific reason, rather than mandates for ‘racial inclusion.’” Scientists agreed that careful genome mapping would reveal the true nature of race.

5. The SNP Map and Beyond

The final years of the Human Genome Project brought to genome mappers the realization that to fully actualize the public health mandates issued by Congress and the federal agencies at the helm of their research, they would have to commit fully to utilization of the federal race category set. It would not be enough to use labels in recruitment and then disambiguate samples. To serve underserved populations and minority communities, and to address racial health disparities, they would have to utilize these particular labels throughout research. Genome Project leaders moved to create a diversity project in this mode.

The SNP Consortium, a partnership between the Genome Project and thirteen major pharmaceutical companies aimed to map one thousandth of humanity’s genetic variation, approximately 300,000 SNPs.18 The Consortium began with Discovery Resource samples “representing several racial groups” a là the federal category set (US Department of Energy 2001), and each company in turn contributed samples and analysis based on this racial model. This meant taking samples that were labeled in terms of a group or individual’s self-identity—connoting anything from ethnicity and biogeography to religion and linguistic difference—and relabeling them as one of the continental groups recognized by Directive No. 15. Once again, Congress’s original charge that such categories not be conflated with biological concepts of race was forgotten, and difference was framed as based in the DNA of said racial groups. Importantly, the Consortium took a labels-on approach in which they publicized samples with a racial affiliation. All data were submitted online using US federal racial classifications (Thorisson and Stein 2003).

In no time, the project was able to map over one million SNPs. Yet genome mappers were even more elated to find that polymorphisms were passed down in clusters of sequence called haplotypes. As Consortium data were racialized, haplotyping was analyzed only in terms of race. Perlegen, for example, claimed that “more than eighty percent of a global human sample can typically be characterized by only three common haplotypes” . . . the “African, Asian, and Caucasian chromosomes” included in their pool (Patil et al. 2001) Genaissance reported significant haplotypic variation in “twenty-one whites, twenty blacks, twenty people of Asian descent, eighteen Latinos, and three American Indians” (Dunham 2001). Study scientists also conducted a comparison between the Consortium’s European and African samples to characterize ancestral population divergence. Following this research, Perlegen was brought on as a primary sequencing center for the new haplotype-mapping project.

The National Human Genome Research Institute initiated the International HapMap Project, the largest ever genome sequencing project dedicated to mapping humanity’s diversity in a way that could lead to cures for underserved groups.19 In an early sampling meeting, planners debated how to go about stratifying global variation. They wanted to use federal race categories for recruitment and sample labeling; however, they were concerned that research labeled as such would be misused. Racial stigmatization was of utmost concern to planners (Murray 2001). Lander told reporters, “We must make sure the information is not used to stigmatize populations. But we have an affirmative responsibility to ensure that what is learned will be useful for all populations. If we shy away and don’t record the data for certain populations, we can’t be sure to serve those populations medically” (Wade 2001).

Project planners settled on a continental sampling approach that would permit future research to investigate racial differences. As Lander explained it: “We know that there are regional similarities. We know that in Africa there is somewhat more variation, because all mankind comes from Africa. And so in order to get some idea whether regional variation mattered, it was useful to pick a site in Africa, a site in Europe, and a site of Asia. If one simply picked with worldwide grid sample, you wouldn’t find out” (Bliss 2012, 59).

For the HapMap Project, which like the Diversity Project rejected grid-sampling in favor of using preexisting classifications, diversity requirements in scientific and ethical terms were believed to be best served with the continental-racial approach. Project leaders appointed scientists with connections to each of the continental recruitment targets to head up the various branches of the project. With community leaders on the ground, they decided that those sampled would be brought into the project with ceremonies and fanfare that held significance for those participants.

In deliberations with international partners, various labels were negotiated according to local demands for representation. For example, the African sample was collected from Yoruba located in Nigeria, and that population requested their samples be labeled as such. Also, the Chinese and Japanese groups that contributed to the Asian portion of the pool negotiated a label that retained that complexity. The project did not include any Indigenous American labels as a result of Native American groups opting out entirely. As Pui Yan Kwok recalled: “[T]hey just did not think that they [the project findings] would benefit them necessarily. Number one, they are giving us something, but we are not giving anything back. And the other is that they don’t like the conclusions that we are going to draw from the data . . . because they have their own traditions about how they came about.” Despite these setbacks, the DNA was managed and distributed in a way that was amenable to racial research.

In 2003, the Human Genome Project was brought to a close. Yet this was a new beginning for genomics with race, with the field as beacon and arbiter of the meaning of race and how to best utilize it in biomedicine. Although the field was now deeply invested in the federal conceptualization, it was generating more “proof” of the veracity and utility of race in its own use of it. Furthermore, the National Institutes of Health now looked to the Genome Project leadership, especially the National Human Genome Research Center’s ELSI branch, to lead its efforts to foster inclusive science and health care, battle health disparities, and create a new precision medicine initiative, one relevant and tailored to people of all backgrounds.

Two initiatives in particular evince this trend. First, the agency issued a policy regarding genetic sampling that required minority community consultation on all studies (National Institutes of Health 2000). This policy followed ELSI recommendations but was aimed at all institutes. Second, the agency issued the Strategic Research Plan to Reduce and Ultimately Eliminate Health Disparities, which provided institutes with $1.3 billion to rationalize research with a greater emphasis on health disparities and genetics (US Department of Health and Human Services 2000). This dual focus sent the message to funders and researchers that the new genomics would thereafter be a primary weapon in the fight against health disparities.20

The National Institutes of Health’s trans-institute plan to address health disparities also funded a number of minority research awards and training programs designed to fast-track minority scientists in their careers and bring about minority community research oversight. Initiatives like the Genomics Short Course for Faculty at Minority Institutions and the ELSI career award further coupled genomic rationalization and racial equality with a focus on people power. Each of the institutes established their own plans to fight health disparities, and this approach to race was also reflected in their plans. The National Human Genome Research Institute’s plan was exemplary in this regard. Nearly all of its initiatives dealt with minority recruitment into decision-making and scientific roles.

6. Post-Human Genome Project Race

In the wake of the Human Genome Project, and the establishment of its successor project, the HapMap Project, “revitalizing” race became a mantra for the field. The National Human Genome Institute embarked upon a campaign of racial research with two of its leaders who also headed up genomics research at the historically Black university Howard University, Georgia Dunston and Charles Rotimi. Howard University’s National Human Genome Center, “a comprehensive resource for genomic research on African Americans and other African Diaspora populations, distinguished by a diverse social context for framing biology as well as the ethical, legal, and social implications of knowledge gained from the human genome project and research on genome variation,”21 had celebrated the transition from Genome Project to HapMap Project with an interdisciplinary conference titled “Human Genome Variation and ‘Race.’” The conference led to a special issue of Nature Genetics that featured the views of Genome Project leaders and other leading genome researchers and experts known for their cutting-edge knowledge on race and genomics. Following this meeting, ELSI made the National Human Genome Center a Center of Excellence in ELSI Research. One of the first ever institutes awarded this status, it became the beneficiary of a multimillion-dollar fund to bring the field’s top research priorities into view.

Dunston and Rotimi moved to create the first race-based genome project: The African American Diversity Project. Like the diversity-mapping projects before this project, they framed the project as a necessary redress to the Eurocentrism that plagued biomedicine broadly speaking and genetic science particularly. They continued the now paradigmatic practice of melding research inclusion with sample diversification and social justice, namely through aid to underserved minority groups. Dunston called for filling the “critical” knowledge gap, stating “there is no substitute for participation” (Goldstein and Weiss 2003). Both warned that without participation, precision medicine would pass people of African descent by. As Rotimi argued, “If you want your clothes to fit, you’d better go to the tailor to be measured” (Goldstein and Weiss 2003; Taylor et al. 2004). In their eyes, African descent meant shared ancestry worthy of a large-scale sequencing project of its own, one that would guarantee therapies tailored to that ancestry. African American advocacy organizations rallied around the project, agreeing that being sequenced would ensure access to genomic therapies.

The African American Diversity Project was soon eclipsed by news that race-based medicine was finally ready to be brought to market. The AHeFT trials were complete. The heart failure drug BiDil had been shown to produce a 43 percent relative one-year mortality decrease in its Black subjects (Taylor et al. 2004). Once again, advocacy groups rallied. In Congressional hearings in 2005, they stood with the drug’s maker Nitromed petitioning law and policymakers to authorize the funding and approval of race-specific drugs.22

As HapMap Project leaders began envisioning their successor project, a whole genome sequencing project that would capitalize on emerging next-generation sequencing technologies, African American Diversity Project leaders met with potential partners in Africa to discuss broadening out the project across the African sub-continent. A private diversity-mapping project sponsored by the National Geographic Society, the National Genographic Project, had met with great success in crowdsourcing samples from all over the world, including from the African groups believed to have haplotypes important to understanding humanity’s first migrations within and outside of Africa despite challenges from a number of Indigenous groups in other parts of the world.23 The National Institutes of Health and Wellcome Trust, the primary sponsors of the Human Genome Project, once again teamed up to support the Human Heredity and Health in Africa Project, a multimillion-dollar project dedicated to understanding gene-environment interactions in people of African descent. Of this project Rotimi charged, “we would like to make sure that whatever benefit that we’re going to accrue from using genomics to understand health and human history that [it] doesn’t go past Africa, as other revolutions have done in the past.”24 The project resulted in being the first official race-based large-scale sequencing project, the fruition of decades of attempts to bring race into the heart of genomics.

7. Conclusion

Since then, there have been more global genome projects made in the image of the late Human Genome Project. The racial schema has permeated these projects, as has the “inclusion means cures” rationale. As a result, the field’s main databases, from which the great majority of studies are done, are racialized by the federal category set and continental groupings more broadly construed. This racial nature leads to deeply embedded assumptions about innate racial differences in biology, and to racial comparison within and across studies.

But the impetus to use race thusly has implications for studies conducted on other samples as well. Genetic clustering software is used all the time to sift through pools of DNA samples, and to separate out certain groups to be used in studies. Population substructure, or the existence of population variation within a sample pool, is an issue for which all researchers must correct. With the given modus operandi, all too often clustering software is used to parse by racial group (Maglo et al. 2016). Researchers separate out people of Asian or African or European descent and stop there.

My research with Aaron Panofsky of UCLA’s Institute for Society and Genetics has shown that, indeed, continental differences are the rubric within which all differences are filtered (Panofsky and Bliss 2017). Even when studies are comparing the DNA within a population that is more akin to a highly localized ethnolinguistic group, researchers and science reporters have a tendency to report results in those racial terms delineated in the federal category set. Thus, this mainstay method of handling human variation leads us back to race.

Researchers have high hopes that their science will be a driver of racial justice. Yet in uncritically applying genomic methods to stratify populations by race, and then linking these groups to all kinds of traits and diseases, race is reified as genetic. To bring us to a truly equitable society, researchers will need to refrain from using racial classifications to stratify DNA samples, even if those categories were used in recruitment.25 They will need to also refrain from using clustering software to parse by continental group. Instead, researchers will need to study variation on a finer level than continental clustering. This will prevent wanton characterization and comparison by race.

To conclude, I would like to affirm the enduring need for the inclusion of racial minorities in biomedical research. Yet, I see a way forward that can avoid these slippages. First, maintain a difference between recruitment labels and analysis labels. Recruitment categories should not be affixed to samples indefinitely, and they should not be reinstated during genomic analysis. Second, use racial categories only in analysis of the biological effects of racism. There is great value in researching gene-environment interactions involved in the experience of racial discrimination. Yet, racial classifications must be reserved for such research.

Notes

1. 1. The Human Genome Project emerged from many fronts, including but not limited to the US Department of Energy, the National Institutes of Health, the Wellcome Sanger Institute, and Genethon. Though they possessed many different notions of sampling, methods, and significance, this chapter shows how US federal agency priorities around racial inclusion moved leaders of the project to assess the project’s implications for race, while inspiring those spearheading new international projects to eventually adopt a racial framework.

2. 2. Interviews were in-person, in-depth, and open-ended. Ethnographic observation included shadowing in workplaces and laboratories, participating in meetings and conferences, and spending informal time outside of workplaces. See Bliss (2012) for a detailed methodology.

3. 3. The first meeting dedicated to policy discussions for the field did not occur until 2003 at Howard University, after the International HapMap Project was well underway. Conference themes were made public in a supplemental issue of Nature Genetics in 2004 (“Genetics for the Human Race” Nature Genetics Supplement 36 [2004]).

4. 4. See Epstein (2007) for a look at the institutional debates that took place in the late 1980s within the National Institutes of Health over how to implement Congress’s federal racial taxonomy.

5. 5. Sequencing centers included the US Department of Energy Joint Genome Institute, Walnut Creek, California; Baylor College of Medicine Human Genome Sequencing Center, Department of Molecular and Human Genetics, Houston, Texas; the Wellcome Trust Sanger Institute, Wellcome Trust Genome Campus, Hinxton, Cambridgeshire, UK; Washington University School of Medicine Genome Sequencing Center, St. Louis, Missouri; and Whitehead Institute/MIT Center for Genome Research, Cambridge, Massachusetts. US Department of Energy (DOE), “Human Genome Project,” in Human Genome Research Sites, http://www.ornl.gov/sci/techresources/Human_Genome/research/centers.shtml.

6. 6. In 1993, Centers for Disease Control and Prevention also drafted the “Use of Race and Ethnicity in Public Health Surveillance” statement: Centers for Disease Control and Prevention, “Use of Race and Ethnicity in Public Health Surveillance Summary of the CDC/ATSDR Workshop,” MMWR 42[RR-10] (June 25, 1993). For a report on Centers for Disease Control and Prevention implementation of OMB categories, see Centers for Disease Control and Prevention, Federal Register 64, no. 46 (Mar. 10, 1999), 11915–20.

7. 7. Planners were blinded by the predominant continental ancestry assumptions that saw admixture in terms of mixture between continental groups. For example, they considered Europeans to be a population apart from Africans, thus they ignored the admixture within people of European or African descent.

8. 8. Human Genome Diversity Project, address delivered by Luca Cavalli-Sforza, Stanford University, to a special meeting of UNESCO (Paris, Sept. 21, 1994).

9. 9. Cavalli-Sforza had long denounced potential antisemitic uses of genetics, but now governing bodies were charging his project with similar racial abuses.

10. 10. Further research that I conducted with sociologist Aaron Panofsky has shown that undefined continental frameworks that correspond to the racial groupings of the US federal category set have predominated in science reports ever since. In publication, researchers slip between the language of “population” and “ancestry” in the broadest continental comparative terms. Thus, even race-neutral language connotes racial meaning.

11. 11. Federal race categories were intended to be representative of groups who shared a sociopolitical history determined by societal conceptions of difference. They were explicitly defined in terms of societal understandings as opposed to biological reality.

12. 12. US Department of Energy, “A DNA Polymorphism Discovery Resource,” Human Genome News 10 (Feb. 1999), http://www.genome.gov/pfv.cfm?pageID=10001552. This statement was later edited for the Discovery Resource webpage to read: “The resource includes samples representative of the genetic diversity found in the U.S. population.”

13. 13. Some of the first projects funded were: Sandra Lee’s “The Ethics of Identifying Race in the New Genetics”; Howard Markel’s “The Stigma of Disease: Implications for Testing”; and David Micklos’s “Digital Image Archive on the American Eugenics Movement” (cf. Stevens, “Racial Meanings and Scientific Methods”).

14. 14. Office of the Press Secretary, The White House, remarks by the president, Prime Minister Tony Blair of England (via satellite), Dr. Francis Collins, Director of the National Human Genome Research Institute, and Dr. Craig Venter, President and Chief Scientific Officer, Celera Genomics Corporation, on the completion of the first survey of the entire Human Genome Project (The East Room, June 26, 2000; 10:19 a.m. EDT).

15. 15. Office of the Press Secretary, The White House, remarks.

16. 16. Study scientists used the word “ethnicity,” yet they were looking for vast continental differences in their clusters.

17. 17. Wood coauthored the Xie, Kim, Wood, and Stein (2001) review designed to establish drug response gene frequencies as ethnically variant.

18. 18. Participating companies included APBiotech, AstraZeneca Group PLC, Aventis, Bayer Group AG, Bristol-Myers Squibb Co., F. Hoffmann-La Roche, Glaxo Wellcome PLC, IBM, Motorola, Novartis AG, Pfizer Inc., Searle, and SmithKline Beecham PLC. Consortium-funded labs included the Whitehead Institute, Sanger Centre, Washington University, and Stanford University. Data were processed at Cold Spring Harbor Laboratory.

19. 19. For a detailed look at the sampling protocols utilized in the transition from the SNP map to the HapMap, see Rajagopalan and Fujimura (2018).

20. 20. This was similarly focus of the Department of Health and Human Services’ Initiative to Eliminate Racial and Ethnic Disparities and Healthy People 2010, which provided broad guidelines to public health agencies, scientists, and medical providers alike.

21. 21. http://www.genomecenter.howard.edu.

22. 22. Nitromed enlisted many other Black health advocates (see Puckrein 2005; Puckrein and Yancy 2005). Research subsequently showed the inadequacy of the race-based approach (see Maglo et al. 2014).

23. 23. Despite protest from several Indigenous and First Peoples organizations, including the United Nations Permanent Forum on Indigenous Issues, numerous polities from all continents agreed to take part. By the project’s close in 2010, over fifty thousand tribes had been sampled.

24. 24. http://www.genome.gov/Pages/Newsroom/Webcasts/Transcript-H3-AfricaPressConference.pdf.

25. 25. Moreover, recruitment uses must be based on an explicit aim to study issues pertaining to race, such as biological effects of racial discrimination.

References