Distributed and Concentrated Strategies in the Sequencing of the Yeast Genome

Abstract

This chapter examines the sequencing of yeast (Saccharomyces cerevisiae), which was conducted by an international collaboration that released its reference genome in 1996. The European side of this collaboration exhibited a distributed model of sequencing, in which yeast biology laboratories made immediate research use of the sequence data they produced. In the USA, by contrast, large-scale sequencing centres were established, with yeast genomics constituting a pilot for the eventual sequencing of the human genome. The European effort involved a heterogenous community of yeast ‘genomicists’ and constituted a distinct model of genomics not subordinated to the whole-genome sequencing of Homo sapiens. Most of the human genome projects that proliferated in the early to mid-1990s instantiated some form of this distributed model, which co-existed with the more concentrated and intensive production of the human reference genome.

John Sulston, the scientist who led the British contribution to the human reference genome sequence, considers that what is now called the Human Genome Project really got started with simpler organisms in 1989. That year, he passed a point of no return in his career that led him to see the sequencing of whole genomes through the scaling-up of technologies and scientific teams as the only way forward. This moment, which Sulston compares to a “prison door” shutting behind him, occurred during the seventh international meeting on the nematode worm Caenorhabditis elegans, held at Cold Spring Harbor Laboratory (CSHL) on the south coast of Long Island Sound, in New York state. There, Sulston and his associates Alan Coulson and Robert Waterston unveiled a physical map of ordered DNA fragments that encompassed the whole genome of C. elegans. This worm, of about one millimetre length, had become both Sulston’s obsession and a widespread model organism in genetics research over the preceding 20 years. In 1983, after tracing all the divisions of C. elegans cells during embryonic and post-embryonic development, Sulston had embarked on assembling the map of its genome, first with Coulson and later in collaboration with Waterston. When James Watson, director of CSHL and Nobel Prize winner for his co-discovery of the structure of DNA, saw the map at the meeting, he exclaimed: “you can’t see it without wanting to sequence it, can you?” (Sulston & Ferry, 2002, pp. 13–14).

Watson was by then combining his long-term directorship of CSHL with a new appointment as associate director for Human Genome Research at the US National Institutes of Health (NIH). His remark propelled a frantic series of meetings in which Sulston, Coulson and Waterston committed to sequence 3% of the worm’s genome, the largest portion of DNA that had been tackled to date. Watson offered to support the operation through the Office for Human Genome Research, which the NIH had established in 1988. Following favourable review of the detailed proposals, the NIH funded the whole sequencing enterprise in the USA—led by Waterston’s team at Washington University in St Louis (WU)—and one-third of Sulston’s work at the Laboratory of Molecular Biology (LMB) in Cambridge (UK). The rest of the funding was provided by the UK Government through its national Human Genome Mapping Project (Chap. 3). This international initiative started in 1989, just months after the CSHL meeting. Three years later, in 1992, three of the 100 million nucleotides of the worm genome had been completed. The sequencing effort was presented as a “pilot system” to test the technologies and feasibility of addressing the human genome, as well as interpreting the resulting data (Sulston et al., 1992, p. 37). The human genome comprises 3 billion DNA nucleotides, and so is about 30 times larger than that of C. elegans.

What is less known—and absent from Sulston’s account—is that C. elegans was one of the drivers in the sequencing of another genome: that of the baker and brewers’ yeast, Saccharomyces cerevisiae. Prior to that fateful 1989 meeting, Waterston had started using Yeast Artificial Chromosomes (YACs) in the physical mapping of the worm. This tool, developed in the 1980s using recombinant DNA techniques, allowed the insertion of foreign genetic material into yeast cell cultures. By using the replication mechanisms of yeast—a single-celled fungus—researchers could multiply (clone) the foreign inserts and obtain enough DNA of the organism they were working with. Waterston inserted the worm DNA fragments that he wanted to map into YACs and then multiplied them, producing a library of C. elegans fragments stored in S. cerevisiae cells. The fragments could then be isolated and their position within the worm genome determined. Given that this procedure yielded large amounts of yeast as well as C. elegans DNA in the cell cultures, Waterston included a project in his NIH grant to sequence S. cerevisiae on top of the worm, to distinguish the DNA of the two species. The project, which ran in parallel with the C. elegans sequencing effort, enabled WU to join an incipient multi-national and multi-institutional initiative to complete the entire yeast genome.

This chapter shows how Waterston’s two-species effort signalled the emergence of a handful of groups that absorbed unprecedented amounts of funding for comprehensively mapping and sequencing whole genomes. Despite these comprehensive efforts starting with microscopic organisms such as yeast and C. elegans, the intention of their sponsors from the outset was to concentrate resources and capacities that these groups could later deploy for the human genome. Apart from WU, the NIH channelled its funding towards another yeast sequencing group at Stanford University (Szymanski et al., 2019, p. 36). In 1993, when the yeast sequencing effort was proceeding apace towards completion, the NIH transformed these two groups into the Genome Sequencing Center at WU and the Genome Technology Center at Stanford. That same year, Sulston left the LMB to become the founding director of the Sanger Institute, an institution that would comprehensively map and sequence the C. elegans, yeast and human genomes with funding from the UK Government and, especially, the British biomedical charity Wellcome Trust (Chap. 4).¹ In the sequencing of S. cerevisiae, WU, Stanford and the Sanger Institute cooperated and competed with smaller institutions from Canada, Japan and the USA, as well as a transnational consortium of laboratories sponsored by the European Commission (EC).

The chapter compares the strategy of concentrating funding and resources in specific groups that subsequently became genome centres, with the approach that the EC undertook in the sequencing of yeast. Unlike the NIH and the Wellcome Trust, the EC avoided channelling funding into just one or a few teams and preferred instead to distribute its support across a wider range of laboratories based in multiple European countries. This partly followed from the agenda of what was then called the European Community—from 1993 the European Union—and the opportunities that a networked yeast genome project presented for fostering political and economic integration among member-states (Parolini, 2018). In contrast with its counterparts in the UK and the USA, the EC did not regard yeast sequencing as a springboard to tackle the larger human genome: it was, rather, a means to encourage and cement cross-European scientific and industrial collaboration, based on the potentialities of yeast as a model and industrial organism. This led the yeast sequencing consortium to be dominated by academic and corporate laboratories that were already investigating S. cerevisiae as a biotechnological object, a brewing instrument, or a model organism for genetics and cell biological research. The consortium also included a group of companies—some of them start-ups arising out of universities and publicly-funded research institutes—that provided sequencing services for S. cerevisiae and other genome projects sponsored by the EC.

The coordinated action of these laboratories created a distributed approach to sequencing that, as we show below, was re-implemented in subsequent EC projects and shared by the majority of national human genome programmes that emerged from the late-1980s onwards (Chap. 3).² We argue that this approach diverges from the canonical history of genomics—and its hourglass representation (Chap. 1)—in that a heterogeneous array of institutions exhibited diverse ways of sequencing DNA and modes of interacting with each other and external bodies. Crucially, these institutions persisted in their operation, without being replaced by a more homogeneous landscape of genome centres. However, the concentrated strategy that the NIH and Wellcome Trust pursued, along with changes in EC policies, increasingly reduced the visibility and scope of this distributed mode of genomics into the 2000s.

1 Out of C. elegans Sequencing

By 1989, the year of the crucial CSHL meeting, Sulston, Coulson and Waterston had established themselves as key drivers of the C. elegans community. As historians have documented, this tiny worm had become a widespread model organism for genetics research in the 1970s (Ankeny, 1997), and Sulston and Coulson’s ‘fingerprinting’ mapping techniques had subsequently emerged as an obligatory passage point for the investigation of C. elegans genes. From the mid-1980s onwards, an increasing number of laboratories sent samples of C. elegans DNA to the LMB that they had previously identified as corresponding to—or located nearby—genes involved in behavioural, developmental or any other biological functions in the worm. Sulston and Coulson would position the DNA samples within their ongoing physical map and report the results back to the laboratories who had sent them (de Chadarevian, 2004; García-Sancho, 2012b).

Knowledge of the chromosome or chromosomal region in which their samples were located enabled the laboratories to progress their research, allowing them to detect and isolate other DNA fragments as part of the genes they were pursuing. Sulston and Coulson, on their side, could refine their maps by adding the samples they received and increasing the overall number of ordered fragments (Fig. 2.1). The samples were initially delivered to Cambridge as cosmid clones: colonies of bacteria that had propagated from a single one in which the genetics laboratories had inserted the DNA they wanted to be mapped. Upon fingerprinting analysis, Sulston and Coulson looked for overlaps between the sequence of the C. elegans DNA contained in the sample and others from cosmids they had already mapped. Overlapping sequences suggested that the corresponding DNA fragments were contiguous in the worm’s genome. After Waterston joined the team in 1985, he screened for additional sequence matches with the larger YACs that he compiled in St Louis.

This strategy of completing the C. elegans map while providing a service to other laboratories shifted significantly when Sulston, Coulson and Waterston embarked upon the sequencing of the worm’s genome. Following their conversation with Watson at CSHL, the three scientists started advocating an approach in which their laboratories at WU and LMB would sequence the whole worm genome on their own initiative, rather than relying on requests and sample deliveries from C. elegans geneticists. This was the same approach that Watson sought to implement for the mapping and sequencing of the human genome, the task he had been set in his new position at the NIH Office. As we show in the next chapter, in 1989 he was finalising an agreement with the US Department of Energy by which this institution and the NIH would contribute three billion dollars towards the completion of the human genome map and sequence between 1990 and 2005 (Chap. 3). This programme—unprecedentedly large in the molecular life sciences in its level of funding and 15-year time horizon—enabled the C. elegans mappers to undertake a comprehensive, whole-genome sequencing operation; one that would not be conditioned by external requests of sequence data.

Fig. 2.1

Left, Alan Coulson beside the physical map of C. elegans, pinned on the wall of the central theatre of Cold Spring Harbor Laboratory in 1989. James Watson looked at the map and proposed that he, John Sulston and Robert Waterston sequence the genome of this nematode worm. Right, the outcome of Sulston and Coulson’s fingerprinting technique: an autoradiograph picture of a gel on which the DNA fragments to be mapped had been run, one in each column, along with marker fragments that were used for reference. Each black spot on the picture corresponds to a sub-fragment into which the original fragments had been fractionated. Since the enzymes used to fractionate the fragments cut at specific sequence sites, a matching pattern of spots in two or more columns—or across different autoradiographs—meant that the fragments overlapped in the genome. Left image courtesy of Barry Honda and retrieved from Jenny Shaw (2014) “Alan Coulson’s science of collaboration”, blog post produced for the Wellcome Library and available at: https://wayback.archive-it.org/16107/20210313022805/http://blog.wellcomelibrary.org/2014/07/alan-coulsons-science-of-collaboration/ (last accessed 7th December 2022). Right image: reproduced from Sulston, S, Mallett, F, Staden, R, Durbin, R, Horsnell, T, & Coulson, A, Software for genome mapping by fingerprinting techniques. Bioinformatics, 1988, 4(1), 125–132: Fig. 1 on p. 126, by permission of Oxford University Press

In Cambridge, Sulston and Coulson approached the Medical Research Council (MRC) as a potential funder of the two-thirds of the worm’s sequencing project that the NIH did not provide for. The MRC is the agency of the UK Government that funds and oversees biomedical research, including the operation of the LMB. In 1989, the same year as the CSHL meeting, the UK Treasury had granted the MRC an extra 11 million pounds to run a national programme to map and sequence the human genome. Yet the UK programme, called the Human Genome Mapping Project, only had guaranteed funding for 3 years compared to the 15 years of its US counterpart and was not committed to whole-genome mapping and sequencing. This led Sulston and Coulson to propose, in their funding application to the MRC, a phased approach that would start by focusing on targeted regions of the worm’s genome, and then develop the efficiency and output of the sequencing techniques, culminating in a “factory style operation”. The end goal was to move beyond the three-year support framework, and comprehensively sequence DNA fragments from mapped cosmid clones and YACs encompassing the entire C. elegans genome, so it would be completed “in a time not longer than 10 years”.³

Waterston’s parallel proposal to the NIH took the intended efficiency and comprehensiveness of such a factory model further. At the same time as the C. elegans operation, his department at WU had hosted another mapping initiative aimed at the yeast S. cerevisiae and led by Maynard Olson (Szymanski et al., 2019, pp. 435ff). Olson had provided yeast DNA to Waterston and other colleagues for the construction of YACs. The availability of this source of DNA, together with the use of YACs in the mapping of C. elegans, led Waterston to include a side project—called Project 2—to sequence S. cerevisiae as well as the worm in his formal funding application to the NIH. From a C. elegans sequencing perspective, this project sought to distinguish between yeast and worm DNA in the YACs. Yet by using Olson’s map, the yeast sequence data could also be assembled and stored, rather than just being discarded as contamination.

By the time Waterston was writing his proposal, Olson had moved from St Louis to a new Department of Molecular Biotechnology at the University of Washington in Seattle. Instead, yeast geneticist Mark Johnston joined Project 2 and more generally became responsible for yeast sequencing at WU after 1992. Olson’s former technician at WU, Linda Riles, started working on the NIH grant and providing clones, including in YACs, whose sequencing would be collaboratively overseen by Johnston and Waterston.⁴ Johnston inherited from Olson the spirit of distributing clones with the mapped yeast DNA fragments to other laboratories that were starting sequencing projects. Among these were the members of the EC consortium and two independent groups of Canadian and Japanese institutions led by McGill University and RIKEN (Rikagaku Kenkyūjo, the Institute of Physical and Chemical Research), respectively.

Waterston’s proposal was funded as part of the NIH contribution to the human genome through the National Center for Human Genome Research. This institution had succeeded the Office for Human Genome Research in 1989 and also sponsored one-third of Sulston and Coulson’s endeavour (Fig. 2.2). Waterston’s funding was for five years. In the case of yeast, his grant aimed to complete the whole of chromosome VIII of this organism “by mid-1994” and sequence “parts” of other chromosomes “totalling 2.5” million nucleotides. It was envisaged that, along with other efforts “in progress or planned worldwide”, the entire S. cerevisiae genome would be determined “by the end of 1995”.⁵ The WU team completed the sequence of chromosome VIII by the target year of 1994 (Fig. 2.3) and led the determination of another full chromosome—XII—that was published in 1997, co-authored with the EC consortium. The WU group also contributed to chromosomes IV and XVI in collaboration with the Sanger Institute, Stanford University, the EC consortium, and McGill and Concordia Universities in Canada, as well as the University of Toronto Hospital for Sick Children.

Fig. 2.2

The cover of a 1992 meeting programme at Cold Spring Harbor featuring John Sulston and Robert Waterston as C. elegans worm cyclists competing with other organisms in a race to finish their genomes. The race is refereed by James Watson, pictured beside the ‘Start’ sign with his hand raised. Courtesy of Cold Spring Harbor Laboratory Archives. The image was introduced to us by Marina Schutz, who analysed it at the workshop “Cooperation and Competition in the Life Sciences”, held in November 2019 at the Ludwig-Maximilian University of Munich. It is also available at: Papers and Correspondence of Sir John Sulston, Wellcome Library, London (UK), file number PP/SUL/A/6/16

Fig. 2.3

Genetic linkage and physical maps that Mark Johnston used to sequence chromosome VIII of the yeast S. cerevisiae. In the genetic linkage map—upper part of the figure, marked as ‘A’—the vertical lines represent relative positions on the chromosome and the numbers denote relative distances. In the physical map—middle of the figure, marked as ‘B’—the horizontal lines represent partially overlapping DNA fragments and the alphanumeric codes denote the samples from which those fragments were obtained, such as a cosmid library or another form of cloning system. The lower part of the figure, marked as ‘C’ corresponds to the data submitted and stored in GenBank, the global DNA sequence database that was established in the USA in 1982. The whole yeast chromosomal sequence was fractionated into various, non-overlapping manageable chunks that are represented as horizontal lines with their submission references listed below them. Johnston et al. (1994: 2080, Figure 1). Reprinted with permission from the American Association for the Advancement of Science

Another element in common between the NIH-led sequencing of yeast and C. elegans was that, in both cases, the effort was coordinated between two groups. Yet in the case of S. cerevisiae, the WU group partnered with a US institution—Stanford University—rather than a British one. While the partnership between WU and the LMB derived from Waterston and Sulston’s collaboration in the prior mapping of C. elegans, with yeast there was not a strong inter-personal or inter-institutional connection beyond Johnston having completed a postdoctoral fellowship at Stanford before moving to St Louis. Johnston’s research at Stanford had explored the genetics of glucose metabolism in S. cerevisiae. His mentor was Ronald Davis, a contributor to the development of the first recombinant DNA techniques at Stanford’s Department of Biochemistry during the mid-to-late 1970s (Yi, 2015, Ch. 2). These techniques, which allow the transfer of genetic material from one organism to another, were used by Johnston to isolate and study the expression of the GAL genes responsible for the synthesis of proteins that process sugars in the yeast cell.

In 1990, the reputed human and microbial geneticist David Botstein moved to Stanford. Botstein had a history of collaboration with this institution, since in 1980 he had co-authored a seminal article with Davis and other colleagues proposing to map the human genome using the restriction enzymes that form the basis of the recombination procedure (the Restriction Fragment Length Polymorphism—RFLP—approach; Botstein et al., 1980; Chap. 1). In the recombination procedure, restriction enzymes cut DNA at specific sequence sites and therefore allow cleavage, transfer and insertion of genetic material from one organism to another where it can be expressed and studied under controlled conditions. Botstein was later appointed vice-president of Genentech, the biotechnology company that Stanford had created to develop and market commercial products derived from the recombinant technologies, such as human insulin expressed in bacteria (Rasmussen, 2014, Ch. 2; Yi, 2015, Chs. 4–5). By the time he moved to Stanford, Botstein had become an enthusiastic advocate of using yeast as both a model to investigate gene function and a system to express recombinant DNA molecules (Botstein & Fink, 1988).⁶

Botstein and Davis were regarded by Watson as representatives of a younger generation of molecular biologists, one that had scientifically and commercially exploited the elucidation of the double helix and other milestones around the structure and function of DNA. This perceived continuity led Watson and others who had been prominent in shaping molecular biology to mobilise a success story around biotechnology and recombinant DNA as a way of retaining their influence and authority. Botstein and Davis were tasked with co-leading a second NIH-funded yeast sequencing group (Szymanski et al., 2019, p. 436). Their Stanford team published the whole sequence of chromosome V in 1997 and was involved with WU in the collaborative sequencing of chromosomes IV and XVI of S. cerevisiae. It also hosted the Saccharomyces Genome Database, a centralised repository that was founded in 1993 and released to the community one year later (Chap. 7).

As the yeast and worm genome endeavours progressed, the St Louis sequencing teams were incorporated into the larger WU Genome Sequencing Center, and Davis and Botstein’s group into the Stanford Genome Technology Center. Both centres extended their remit to contribute to the NIH arm of the (by then fully inaugurated) Human Genome Project of the USA: the one at WU via large-scale sequencing and the one at Stanford through the development of advanced instrumentation (Chap. 3). In 1993, the same year in which these US centres were founded, Sulston became director of the Sanger Institute and created specific divisions for the sequencing of the human and yeast genomes, as well as C. elegans (Chap. 4). A common feature of these three centres was their unprecedented level of funding—considerably higher than the groups from which they derived—and the conditioning of this support on an absolute prioritisation of the genome efforts, rather than them using the sequence data for research into human, yeast or C. elegans biology. This was the fundamental difference between them and the laboratories involved in the yeast sequencing project sponsored by the EC.

2 The Distributed Strategy of the European Commission

The EC initiative was called the Yeast Genome Sequencing Project (YGSP) and started in 1989. In common with the St Louis effort—and later with Stanford and the Sanger Institute—it had the objective of producing a full reference sequence of S. cerevisiae that would encompass its whole genome and function as a representation of this type of yeast. Yet the strategy for achieving this common goal markedly differed between the EC, and the US and British projects. While Stanford, WU and the Sanger Institute comprehensively sequenced entire yeast chromosomes with a view to applying the expertise that they cultivated (and the technologies that they refined) through this to the human genome, the EC pursued the YGSP with a distributed approach.

This distributed approach aligned with the broader political and economic agenda of the EC, the executive arm of the European Community. It resulted in the EC funding being spread across a larger number and variety of both public and private institutions, contrary to the concentration of resources in the WU and Stanford teams, and later the Sanger Institute. Furthermore, the sequencing work undertaken by the YGSP participants was more immediately motivated by research objectives that went beyond producing a reference sequence of S. cerevisiae. Within the YGSP consortium, academic institutions conducted genetics, biochemical and cell biological research with the sequence data, while the companies and corporate laboratories used their experience on the project to foster their commercial activities and expand their customer bases.

The YGSP was part of the EC’s Second Framework Programme for Research and Technological Development. By the time this programme was launched in 1987, the EC—and the European Community more generally—were pushing for stronger political and economic integration among their member-states. This process had gained new impetus with the adoption of the Single European Act—which also came into effect in 1987—and led to the development of policies that would crystallise in the establishment of the European Single Market. Research and development were regarded as key means towards economic convergence, with the EC conceived as the institution that would drive transnational cooperation in science and technology. The funds from the framework programmes were channelled through more specific schemes and, right from the start, biotechnology was identified as an emergent and key area to align scientific and economic agendas across Europe (Bud, 1993, Chs. 7 and 9; see also Bud, 2010, p. S18; Cantley, 1995). The Biotechnology Action Programme (BAP; 1985–1989) and later the Biotechnology Research for Innovation, Development and Growth in Europe (BRIDGE programme; 1990–1993) supported the first years of the YGSP.

Drawing on this political architecture, the sequencing of the yeast genome represented an opportunity to build international scientific networks and contribute to economic development within the increasingly integrated European market. The brewing, pharmaceutical and biotechnology industries in the European Community were, from the outset, considered players in both the sequencing effort—the consortium included corporate participants such as Carlsberg Laboratory and Pharmacia Biotech—and the Yeast Industrial Platform, a group of member-state companies to which the EC provided privileged access to the sequence data (Parolini, 2018).

As project coordinator, André Goffeau was responsible for embedding these principles into the day-to-day running of the YGSP. His dual status of yeast geneticist at the Catholic University of Louvain and civil servant at Directorate-General XII—the department of the EC that oversees research, science and technology—enabled him to press for a yeast sequencing operation as part of the transnational projects included in the framework programmes. With Watson, Goffeau shared a tradition of working with standardised organisms and the vision of a reference sequence as a valuable resource for his research community. From the 1950s onwards, Watson and other molecular biologists had adopted model organisms such as phage viruses or C. elegans as “exemplars” to study different life properties. Similarly, yeast geneticists, biochemists and cell biologists had bred and disseminated a specific strain of S. cerevisiae—named S288C—to conduct their experiments and other explorations of yeast biology, including genome mapping (Strasser & de Chadarevian, 2011; Szymanski et al., 2019, pp. 434ff).⁷

Yet the strategies that Watson and Goffeau imprinted into the NIH and EC yeast genome programmes differed substantially. The latter, in line with EC policy, regarded sequencing as a collective effort to be distributed among the research community, rather than assigned to a reduced number of groups. This resulted in a motley grouping of yeast biologists conducting the various aspects of the sequencing process in the YGSP (Fig. 2.4), a much broader community than the specialist sequencing personnel based in the NIH-funded teams and, later, the genome centres. The distinct communities that coalesced around each programme became the very different kinds of genomicists that sent yeast sequencing on disparate trajectories: in the USA as an antecedent to the Human Genome Project; and as a more species-specific initiative in Europe that was closer to the research requirements of academic and industrial laboratories working with S. cerevisiae.

Fig. 2.4

Members of the consortium of the European Commission, along with representatives from Washington University, Stanford University, McGill University and the Sanger Institute, at the final conference of the Yeast Genome Sequencing Project, held in Trieste (25th–28th September 1996). They are wearing specially-designed shirts illustrating the parts of the yeast genome they worked on. André Goffeau is the third on the left and the picture features the following chromosome coordinators: Hervé Tettelin (far-left, chromosome VII); Mark Johnston (second-from-left, chromosome VIII and co-coordinator of chromosomes IV, XII and XVI); Bernard Dujon (behind Goffeau, chromosomes XI and XV); Horst Feldmann (middle-back, chromosome II); Ron Davis (third-from-right, chromosome V and co-coordinator of chromosomes IV and XVI); Howard Bussey (second-from-right, chromosome I and co-coordinator of chromosome XVI at McGill University); Bart Barrell (far-right, chromosomes IX and XIII and co-coordinator of chromosomes IV and XVI at the Sanger Institute). Agnès Thierry (middle-front), a member of Dujon’s group, is also included in the picture. Photograph courtesy of Karl Kleine

Because of this, the membership of the EC consortium was always heterogeneous and included a large number of institutions. The YGSP started with a three-year pilot phase aimed at chromosome III of S. cerevisiae and led by Stephen Oliver, a microbiologist from what was then called the University of Manchester Institute of Science and Technology (UMIST).⁸ This phase trialled the distributed model and used mapped yeast clones—from the AB972 sub-strain of S288C—distributed by Olson and Riles at WU. The sequence of that chromosome was published in 1992 and involved 35 institutions from eleven different European countries, plus Kobe University in Japan and the New Jersey Medical School in the USA, which were external members of the EC consortium. The pilot phase re-affirmed in outline the basic distributed model of sequencing, though indicated that scale-ups in sequencing output would still be needed in the main phase of the project. An unexpected empirical outcome was the revelation of scores of genes—a majority of those found on the chromosome—that were previously unknown to yeast biologists and would require extensive functional analysis to be characterised (Oliver et al., 1992; Vassarotti & Goffeau, 1992).

The full genome of the species was completed in 1996 and marked by a special issue of Nature that was published one year later, entitled The yeast genome directory. It reported on chromosomes led by the EC effort that had not yet appeared in the literature, as well as those led by Stanford, WU, McGill University and the Sanger Institute: nine out of the sixteen chromosomes of S. cerevisiae in all. By then, membership of the EC consortium had grown to 82 institutions. This consortium coordinated ten of the sixteen chromosomes of yeast—three of them in collaboration with American institutions and the Sanger Institute. The nucleotides determined by the EC consortium represented almost 55% of the overall sequence produced (Parolini, 2018, Figure 1; see also Appendix A at the end of Parolini’s article for a full list of EC laboratories) (Table 2.1).

Table 2.1 An outline of the distribution of work among the different institutions involved in completing the S. cerevisiae genome, indicating the size of each chromosome, institutions participating in the sequencing work, scientist(s) coordinating the operation and year in which the sequence was published

Apart from Goffeau and Oliver’s home institutions, another important nucleus of the EC programme was Munich. This city, which like Louvain and Manchester is furnished with a longstanding brewing tradition, had developed an institutional architecture around biomedical research that resulted in different types of laboratories playing crucial and complementary roles in the YGSP. Due to the impracticality of individually reviewing all of the 82 institutions involved in this project here, we focus on the contributors from Munich as representative of the institutional diversity within the EC consortium. By looking at the connections that these Munich-based laboratories deployed both within and across the YGSP, we also reveal how the EC consortium members interacted between themselves and with the institutions—genome centres and smaller teams—that coordinated the sequencing of other yeast chromosomes.⁹

One leading figure from Munich in the YGSP was Horst Feldmann. A chemist by training, Feldmann was a principal investigator at the Institute of Physiological Chemistry, a research institution affiliated to the Ludwig-Maximilian University of Munich (LMU, the largest university in this city) and the Klinikum of the LMU, a teaching hospital. Since 1968, Feldmann had led his own group using yeast as a model to investigate the genes that control the activity of transfer RNA (tRNA). tRNA molecules are involved in the process of translation by which proteins are synthesised in the cell. They interact with messenger RNA (mRNA) molecules produced through the expression of DNA sequence, with each kind of tRNA bringing a particular amino acid, thereby adding to the growing chain that will make up the protein, as specified by the original DNA that was expressed.

Feldmann had learned sequencing methods during a short visit to the LMB, the institution where Sulston and Coulson were based (Feldmann, 2008, p. 291). The laboratory that Feldmann visited was led by Frederick Sanger, the inventor of the first protein, RNA and DNA sequencing techniques and the scientist after whom the Sanger Institute was named (García-Sancho, 2010). At the time of Feldmann’s visit, the C. elegans mapping project had not yet begun and Coulson was starting his career as a technician in Sanger’s group. Feldmann’s mentor in Cambridge was another Sanger technician, Bart Barrell, who would later run the yeast sequencing operation at the Sanger Institute (Chap. 4).

By the time the YGSP started in 1989, Feldmann’s research was focused on specific genomic regions involved in the distribution of tRNA genes on the yeast chromosomes. These regions, called Ty elements, are found in multiple copies in the genome, as they can create new copies of themselves and jump from one area of the chromosome into others; they are structurally and functionally related to retrotransposons and retroviruses.¹⁰

Feldmann saw the emerging yeast genome initiative as an opportunity to exploit his sequencing expertise and, at the same time, use both the EC funding and sequence data to further his investigation of Ty elements. He joined the consortium from the start and, apart from being involved in the chromosome III pilot effort, led the completion of chromosome II in 1994, coordinating other consortium members. During this time, Feldmann’s team combined the YGSP work with comparison of Ty sequences and analysis of the implications of their conservation: the preservation of particular sequences across different evolutionary lineages—including those featuring yeast species—deriving from a common ancestor (Feldmann, 2008; Stucka et al., 1992; Fig. 2.5). Their detailed knowledge of the Ty regions proved essential in the subsequent completion of other yeast chromosomes, where Feldmann and his collaborators offered crucial intelligence on repetitive patterns in the sequence data to other institutions with less specific expertise.¹¹

Fig. 2.5

Top-left, Horst Feldmann’s group at Ludwig-Maximilian University of Munich in the late-1980s, including Gertrud Mannhaupt (far-left), Rolf Stucka (behind the line) and Christa Schwarzlose (second-from-right), all of them involved in the sequencing of chromosome II of S. cerevisiae. Other groupmembers depicted are Hans Lochmüller, Susanne Mitzel and Robert Krieg (second, third and fourth-from-left), as well as Uschi Obermeier (far-right). Bottom-left, Horst Domdey in the mid-to-late 1980s and, right, a laboratory during the early years of Genzentrum. Sources: Top-left picture, Feldmann (2008, p. 300), copyright (2008), with permission from Elsevier; bottom-left and right pictures: E. L. Winnacker and other authors (undated): Laboratory of Molecular Biology and Biochemistry: München / Martinsried, pp. 8 and 4. Reproduced with permission from Genzentrum and available at Hoechst Archives, Frankfurt, file number H0049176

Another prolific contributor of yeast sequences—and participant in the chromosome II effort with Feldmann’s group—was Genzentrum, an institution co-managed by LMU and the Max Planck Institute for Biochemistry (MPIB, also based in Munich). Genzentrum had been established in 1984 as part of a network of centres with which the German Federal Government sought to foster research using the fledging recombinant DNA techniques: there were other gene centres in Berlin, Cologne and Heidelberg. The Munich Genzentrum was headed by Ernst-Ludwig Winnacker, one of the pioneers of molecular biology in Germany, and it incorporated a set of shared facilities in which scientists from LMU and MPIB could access recombinant DNA and sequencing techniques. On top of this, Genzentrum was equipped with a suite of laboratories in which early career researchers could start their trajectory towards becoming principal investigators.¹²

One of these early career researchers, Horst Domdey, became heavily involved in the sequencing of yeast. Like Feldmann, he combined YGSP work with research on the genetics of S. cerevisiae; in this case, the transcription of DNA into mRNA (Fig. 2.5). Yet given the availability of advanced technological facilities at Genzentrum, Domdey also embarked on sequencing work supported by other EC schemes, namely the Human Genome Analysis Programme (see below).¹³

Unlike Feldmann’s laboratory, in which manual sequencing methods were used, Genzentrum was equipped with state-of-the-art automated instruments that were starting to be marketed by the companies DuPont and Applied Biosystems (García-Sancho, 2012a, Chs. 5–6). The funding from the EC genome programmes provided crucial support to the day-to-day running of these sequencing machines, which were used by other Munich-based researchers working on immunology and animal genetics.¹⁴ In contrast with the Sanger Institute and the genome centres at Stanford and WU, the EC funds did not fully cover the sequencing operations at Genzentrum. Due to this, the facilities needed to combine sequencing work for the EC programmes with a service role supporting research grants undertaken by Genzentrum’s early-career investigators, as well as other scientists at LMU and MPIB.

In 1994, Genzentrum created a start-up company, MediGene, to both commercialise medical products derived from their research and conduct on-demand sequencing work. Brigitte Obermaier, who led the genome analysis team in charge of YGSP assignments at Domdey’s group, became head of sequencing services at this company. MediGene was one among a set of mainly German firms created out of academic research that conducted contract sequencing work for other institutions or concerted genome projects (García-Sancho, Lowe, et al., 2022; Zeller, 2001). They were especially active during the YGSP and other subsequent EC genome programmes: thirteen different companies offering sequencing services were involved as co-authors in The yeast genome directory. MediGene participated in the completion of two different chromosomes of S. cerevisiae, with these sequencing operations being the company’s most profitable line of business.¹⁵

MediGene, along with Genzentrum and Feldmann’s group, were representative of what we have called elsewhere the network model of genomics: a versatile and heterogeneous array of institutions that exhibited different motivations to produce DNA sequence data (García-Sancho, Lowe, et al., 2022). This heterogeneity gave the EC consortium members flexibility, and encouraged an ability to adjust to changing circumstances. Within the consortium, scientists and institutions could produce large amounts of sequence data for various users—e.g. the customers of MediGene—or behave more like a traditional life sciences laboratory and use the sequences they determined for specific research purposes: e.g. Domdey and Feldmann’s groups. Even within the same laboratory, the sequences were often contributed to the YGSP after being used for more immediate research work. By contrast, the genome centres deriving from WU, Stanford and Sulston’s group only practised one model of DNA sequencing that was more distal to the final user and led to the production of large amounts of data without advanced concrete knowledge of what its purpose and destination would be.

At the core of the EC network, another Munich-based institution—the Martinsried Institute for Protein Sequences (MIPS)—compiled the various sequencing results and assembled them into full chromosomes and later a reference genome. MIPS also played a crucial role in assessing the quality of the sequences and inferring biological features from the DNA data. This institution was located in the same campus as Genzentrum, the MPIB and other biomedical laboratories of LMU in Martinsried, a suburb in the south-west of Munich.

MIPS had originated in the late 1980s as a unit of the MPIB Department of Protein Chemistry. This department was particularly strong in the determination of protein sequences, whose techniques had preceded the emergence of DNA sequencing. Pehr Victor Edman, a key figure in the development and automation of protein techniques during the 1950s and 1960s (García-Sancho, 2010, pp. 284ff), had moved to Munich in 1972 and finished his career there.

The first objective of MIPS was to harmonise and unify the different protein sequence databases operating in Europe, the USA and Japan. Yet its director, Werner Mewes, saw the sequencing of yeast as an opportunity to extend to DNA MIPS’s expertise in data handling, analysis and standardisation.¹⁶ Unexpectedly, MIPS became the institution chosen by Goffeau to channel the sequencing results produced by the YGSP. He and other EC administrators engineered a funding system in which all the institutions involved in the S. cerevisiae consortium were incentivised to swiftly produce and submit their sequences to MIPS. MIPS scientists checked the accuracy and quality of the sequences before assembling them into chromosomes and, eventually, a whole genome. They also made the sequences public after a period of 6 to 12 months in which, according to the terms of the EC contract, the sequencing laboratories were entitled to exclusive exploitation of the data (Joly & Mangematin, 1998).

The embargo period was another difference between the EC consortium and the genome centres, which made their yeast sequence data immediately available.¹⁷ This further reflects the contrasting philosophies underlying the distributed and concentrated strategies and what the adoption of one or the other required in terms of support and organisation. The concentration of funding in WU, Stanford and the Sanger Institute meant that these institutions could exclusively focus on producing the sequence without any other financial needs that would require some kind of diversification. Their designation as genome centres emphasised this exclusivity of sequence production and differentiated them from other institutions supported by the NIH, the MRC and the Wellcome Trust. By contrast, the distribution of the YGSP budget among a much larger number of institutions resulted in these institutions having to combine the EC support with other sources of funding. One way of achieving this was using the sequence data that each laboratory produced as a springboard that would ease the award of either other contracts—especially in the case of the sequencing companies—or research grants exploring different aspects of yeast biology. As a yeast biologist himself, Goffeau recognised this necessity and protected the competitive advantage of the sequencing institutions through the exclusive exploitation window. At the same time this model was being implemented, other scientists and EC administrators attempted to export the distributed strategy to the sequencing of other organisms.

3 Distributed Sequencing and Larger Genomes

As the 1990s progressed, the EC extended its sequencing programmes to the genomes of other organisms of interest to science and industry, such as the fruit fly Drosophila melanogaster, the bacterium Bacillus subtilis and the plant Arabidopsis thaliana. All three of these species were, by then, model organisms and B. subtilis had been used extensively as a biotechnological cell factory for the production of multiple chemicals. The consortia that the EC created for those sequencing efforts shared the features of the one it had established for the YGSP: the membership was as inclusive as possible and sought to foster cooperation among member-states. As with the YGSP, the full sequencing of those organisms involved cooperation and competition between the EC-sponsored laboratories and other institutions, mainly in the USA and Japan.

The EC project to sequence the genome of A. thaliana began in 1993, with eighteen institutions concentrating on two chromosomes: 4 and 5. Like S. cerevisiae, A. thaliana has an economically-sized genome comprising five chromosomes. It has, though, over ten times the number of nucleotides as yeast. This European effort was later joined by separate initiatives in France, Japan and the USA.¹⁸ These became formally coordinated in 1996 with the creation of the Arabidopsis Genome Initiative, which was completed in 2000. The active participation of researchers specialising in Arabidopsis biology within the EC consortium enabled them to become crucial actors in the sequencing of genome regions that were difficult to tackle with existing large-scale sequencing methods, a role that had been played in the YGSP by Feldmann’s team and other groups with expert knowledge of yeast biology and genetics.¹⁹

Another similarity with the YGSP was that the Arabidopsis genome was parcelled out, both to the different parts of the international collaboration and to the individual participating laboratories in the European network. In Europe, MIPS once again played a leading role as an informatics centre, reviewing the quality of the sequences submitted by the EC-sponsored laboratories, assembling parts of the genome and analysing it.²⁰ Additionally, many of the same sequencing companies that had contributed significant portions of the European sequence in the YGSP came on board in the main phase of the EC Arabidopsis project in 1996, following the conclusion of the pilot begun in 1993.²¹ Even within the European distributed model, however, a move towards concentration of larger sequencing capacities into a few institutions was evident.

Arabidopsis has been described as the “botanical Drosophila”, and the completed genome sequence—the first for a plant—augmented its more general status as a model organism for plant biology. Arising out of the project to sequence the reference genome, The Arabidopsis Information Resource became an altogether more all-encompassing data infrastructure with ambitions far beyond being a mere repository of DNA sequence and other molecular data. Its promoters conceived it as a means of providing a common basis for representing the species as a whole through the integration of multiple different kinds of data and knowledge, and from this to serve as a platform for the exploration of less well-catalogued species (Leonelli, 2007).

Genome sequencing of D. melanogaster was performed by the European Drosophila Genome Project, a consortium comprising ten laboratories that the EC began to fund in 1997.²² This built on previous EC-supported efforts to physically map the Drosophila genome from 1988. As well as smaller centres conducting the sequencing, the Sanger Institute performed an analogous role to MIPS in the yeast and Arabidopsis projects, assessing and assembling the sequence submitted to it by the participating laboratories.²³ Like the other projects mentioned above, Europe joined forces with American institutions. These included the publicly-funded Berkeley Drosophila Genome Project, another NIH-sponsored genome centre at Baylor College of Medicine and FlyBase, a genetics and genomics database with antecedents dating back to the 1930s, but that was made available on the internet in 1992.

An important commercial player involved was Celera Genomics. This company derived from prior partly-charitable and partly-corporate sequencing efforts led by Craig Venter (Adams et al., 2000; Drysdale & Bayraktaroglu, 2000).²⁴ Celera was founded in 1998 and began contributing to the Drosophila project shortly after. It performed the bulk of the sequencing that resulted in the completion of the reference genome in 2000 (Dove, 2000). Seeking as it did to become a leader in the nascent field of bioinformation, the company used the fly initiative as a testbed for its advanced sequencing and informatics pipelines. As we see in subsequent chapters, Celera’s involvement in genomics was far more extensive than its most famous role in the sequencing of the human genome and concomitant controversy about commercially protecting and restricting access to the resulting data (Chap. 4). One of the key innovations it generated through the Drosophila sequencing project was in the annotation of the genome.

We discuss annotation in depth later in the book (Chap. 6), but it is worth noting here the community-based annotation approach that the Drosophila genomicists pioneered: the so-called ‘jamboree’. The jamboree was an 11-day event held at Celera headquarters in Maryland in November 1999. Over 40 researchers from across the world—mainly from the publicly-funded Drosophila projects—converged there. The aim was to join together the biological expertise of the Drosophila community with computer scientists to identify the genes and other key landmarks in the masses of sequence data produced by Celera (Pennisi, 2000). This would give the fly researchers access to the sequence data and Celera an idea of how they could add value to the data they were producing, a key consideration for their commercial strategy.

From this jamboree event, a model of community annotation was created that informed Celera’s later analysis of the human genome in conjunction with medical geneticists (García-Sancho, Leng, et al., 2022) and that served as inspiration for the annotation of the pig genome (Chaps. 5 and 6). This model was, again, based on combining the expertise of a large-scale sequencing institution—Celera—and smaller laboratories with specialist knowledge in the target genomes, many of whom were funded through the distributed model of the EC. As with the yeast and Arabidopsis genome projects, the intelligence provided by the specialist laboratories proved crucial to adding value to Celera’s sequences and for interpreting specific, biologically-relevant regions of the resulting reference genomes.

In spite of these contributions, the EC’s distributed strategy became increasingly challenged with the emergence of concerted efforts aimed at larger and more complex genomes. With the Human Genome Project in the USA underway since 1990, European scientists and EC administrators debated the effectiveness of collaborative consortia and a networked sequencing operation for more ambitious targets. Some defended the advantages of involving the community in the production of the sequence, as had been the case in the YGSP and contemporary initiatives. Others, however, highlighted the difficulties of recruiting a sufficient number of laboratories to sequence exponentially larger genomes and, especially, coordinating their activity and outputs.²⁵

Adopting a more concentrated model was complicated by the EC’s political need to support multiple institutions across the continent and the scarcity of large-enough institutions that could become sequencing centres for all of the European Union. The Sanger Institute, established in 1993, was a sizeable genome centre based in Europe, but it was already supported in its day-to-day operations by two UK bodies—the MRC and the Wellcome Trust—which gave it some independence from the EC approach. Another potential candidate, the European Molecular Biology Laboratory in Heidelberg (EMBL), never wanted to become a genome centre despite being equipped with advanced technology and expertise, including a centralised database to store DNA sequences (García-Sancho, 2011).²⁶ This lack of orientation towards large-scale genomics resulted in the EMBL participating as a standard sequencing laboratory in the EC consortia and MIPS being selected over this institution as the informatics and data assembly coordinator of the genome projects.

One way in which the EC oriented its operation to larger genomes was through programmes that did not seek to sequence them in full. Examples of this were the Pig Gene Mapping Project (PiGMaP, see Chap. 5), and the Human Genome Analysis Programme (HGAP). The latter started in 1990 and involved the creation of a consortium of human and medical genetics laboratories from different member-states. Rather than determining the full sequence of the human genome—which is almost 300 times larger than the yeast genome—the HGAP laboratories sought to refine existing linkage and physical maps. This involved narrowing down the location of genes or gene markers connected to the research interests of the groups forming the consortium. For this, the consortium members used a technique called complementary DNA (cDNA) sequencing that, rather than tackling the whole human genome, yielded only the parts involved in the synthesis of mRNA, a key intermediary in the production of proteins in the cell. The HGAP laboratories divided the genome into different regions—normally connected to the genes and proteins they were working on—and formed groups devoted to mapping and cDNA sequencing (Chap. 3).

Despite only targeting specific genome regions, these groups required enhanced sequencing capacities. This resulted in Genzentrum in Munich and other institutions with advanced instrumentation—often consisting of centralised, shared equipment—becoming especially active in HGAP. MediGene, Genzentrum’s start-up, was split into two companies, with Obermaier’s sequencing services arm becoming an independent brand called MediGenomix. In 1998, MediGenomix joined with other German sequencing firms involved in the YGSP—AGOWA, GATC and QIAGEN—as well as Biomax Informatics, to form the Gene Alliance, a consortium that aimed to pool their sequencing capacities and capabilities to secure contracts for genome sequencing projects.

The Alliance was formed against the backdrop of the greater intensification and centralisation of whole-genome sequencing efforts propelled, to a large extent, by the emergence of Celera in 1998. Sequencing was becoming larger-scale and higher-throughput. The corporate alliance model was an attempt to keep up in terms of capacity when acting jointly, while enabling individual firms to retain their independence and specific expertise. In addition to their individual capacities, by coming together the Alliance could purchase new high-throughput sequencing machines (capillary sequencers) at a lower unit cost than was available for a smaller order.²⁷

A change in the model and direction of funding at the EC level made it more difficult for individual companies like these to operate alone. Three changes in the years around the turn of the millennium were particularly significant. One was a general shift in research policy, namely a lessened emphasis on establishing wide networks of collaborators, something that had opened the door to smaller laboratories and companies. The drive towards European integration was diverted instead towards the building up of large and well-resourced “centres of excellence” that could enable Europe to compete at a global level across a variety of scientific fields.²⁸ The second change was more specific to genomics, with a shift towards funding functional genomics research (Chap. 7), rather than whole-genome mapping and sequencing (Desaintes, 2008; Gannon, 2000). This removed a potential market from the small sequencing companies that had been able to invest and grow through projects such as the YGSP and Arabidopsis genome sequencing. Finally, the mode of funding also changed, with the removal of the payment-per-nucleotide system that had characterised the earlier genome programmes. Participants were instead paid for labour and materials, favouring non-profit research institutes and universities.²⁹

However, the private sequencing companies based chiefly in Germany had built up a customer base beyond EC-funded projects. Through the Gene Alliance, they were able to build on their experience of working with other laboratories and private firms needing sequencing services. In addition to forming a collaboration in 1999 with Genome Pharmaceuticals Corporation around agrogenomics and pharmacogenomics—thus coupling the Alliance’s sequencing capacity to the Corporation’s functional genomics expertise—they were contracted to sequence the whole genomes of two species: the bacterium Chlamydia pneumoniae and the fungus Aspergillus niger.³⁰

The project to sequence C. pneumoniae derived from an agreement between the Alliance and the German pharmaceutical company Byk Gulden (a subsidiary of the chemical firm Altana) in 1998. C. pneumoniae is a bacterium that causes pneumonia and has been implicated in other diseases such as atherosclerosis. The full sequence of its genome was completed in nine months.³¹ The Alliance operated along similar lines to the yeast and Arabidopsis sequencing projects, a model that the companies were familiar with. Various tasks such as library preparation were divided out among the companies, the sequencing itself was parcelled out to the four non-bioinformatics firms, while Biomax had a role analogous to MIPS in the EC-sponsored projects. A key advantage of the Alliance was that the allotting of work could benefit from the different expertise and business models of its members, with tasks also arranged to enable parallel projects to be managed by individual companies and the Alliance as a whole, alongside their more regular operations. Obermaier of MediGenomix has described the Alliance’s model as “interactive”, with monitoring and evaluation of the ongoing sequencing conducted by Biomax, leading to the identification of regions of lower quality and coverage requiring additional work.³²

In 2000, the Alliance was contracted by the research arm of the large Dutch company DSM (which operates in multiple fields, including nutrition and healthcare products) to sequence the genome of Aspergillus niger, a fungus and species of mould that the company used to produce enzymes and organic acids. The bioinformatics capacity of the consortium was just as crucial as its sequencing output to securing the contract, with the ability of Biomax to fully annotate this genome being a key attraction for DSM.³³ QIAGEN took the lead for the Alliance, and the annotated genome sequence was announced in 2007 (Pel et al., 2007).

By then, the Gene Alliance itself was no longer taking on new business. They found that they were increasingly unable to compete on price with the ever-larger centres in east Asia, North America and Europe, to win whole-genome sequencing contracts. Additionally, the increasing availability of next-generation sequencing machines fostered a shift in the business models of the companies. They enabled them to develop new services and markets based on quick, often overnight, sequencing for research in academia and industry. Without the centripetal force of large contracts through the Alliance, the trajectories of the companies diverged. They had always had to differentiate from each other, as they remained competitors in the same market and region. This was reinforced when the countervailing tendencies that were keeping them aligned in a network diminished.³⁴ With the demise of the Alliance, one of the last vestiges of the European distributed strategy of genomics waned.

The models of the Gene Alliance and HGAP co-existed with the genome centres that proliferated in the USA, UK and other countries during the 1990s. Apart from being single institutions rather than groups or consortia, these genome centres were the executive arm of initiatives that sought to produce a reference sequence of the whole human genome. The publication of that reference sequence between 2001 and 2004 led to the genome centres being identified with a single, coherent and unified ‘Human Genome Project’. Since they were not part of this whole reference sequence effort, the HGAP and work of the Gene Alliance were largely forgotten and excluded from the success narrative of genomics.

Our shift from human to non-human genome projects and addressing of the distributed programmes that the EC sponsored throughout the 1990s has enabled us to present other historical configurations of genomics, beyond the success narrative. As the experience of the HGAP shows, this wider diversity of institutions, approaches and genomicists also applied to human genomics. In the next chapter, we further explore this broader history of human genomics by showing that most of the national human genome projects that proliferated worldwide in the 1990s adopted the distributed strategy of the EC and the Gene Alliance during their early years. We also document the factors through which the concentrated model became identified in the public imagination with a single Human Genome Project, which funnelled the diversity of approaches to human genomics into just one.