Search tips
Search criteria 


Logo of wtpaEurope PMCEurope PMC Funders GroupSubmit a Manuscript
Nat Rev Genet. Author manuscript; available in PMC 2009 November 1.
Published in final edited form as:
PMCID: PMC2672783

Data Sharing in Genomics – Re-shaping Scientific Practice


Funding bodies have recently introduced a requirement that data sharing must be a consideration of all funding applications in genomics. Like all new developments this condition has had an impact on scientific practice, particularly in the area of publishing and in the conduct of research. We discuss the challenges that must be addressed if the full benefits of data sharing, as envisaged by funders, are to be realised.

The field of genomics is regarded as a leader in the development of infrastructure, resources and policies that promote data sharing.1 Examples include the Human Genome Project and the HapMap project — which promote the sharing of sequence data — and the more recent data-sharing structures for genome wide association studies (GWAS), such as dbGaP and the European Genotyping Archive 2 Rapid developments in genomics are widely promoted as being dependent upon such resources, which can be accessed by many researchers for different research uses. They are regarded by many as testimony to the success of the principle of open access. In addition, all of the large funding bodies now make data sharing a requirement of support for all projects, including all hypothesis-driven projects, whose primary purpose is to focus on a specific research question rather than to create data to be used by others. The rationale for these policies is that science and creativity are furthered by access to openly available data, and that data created by publicly funded bodies should be freely available in the research community. While these policies are still in their infancy, their impact is starting to be felt on the planning, execution, and oversight of genomics research, and on the way in which results are disseminated.

Through our empirical work with scientists in the field,3 we have identified some key areas of scientific practice that are being affected by these policies. In this paper we discuss these four areas: the difficulties of acknowledging individual contributions to the generation of data; the way that these policies change the responsibilities towards participants; the implications that this has for maintaining public trust; and the new mechanisms that have been developed for oversight of access to data. These important issues illustrate the tensions that data-sharing policies create for researchers, who must fulfil the requirements of funding bodies, while also protecting research participants and their own career development. Failure to understand these particular tensions and the effect of these policies on scientific practice may have a detrimental impact on global good will and trust in genomics research and on the development of sustainable data sharing practices. Consideration of these issues is timely, as the effects of data sharing policies are starting to be visible and understood, but are also being re-examined, as in the recent case where genotypic data were withdrawn from internet access by the NIH and the Wellcome Trust.4-6

Changes in scientific practice

The data sharing policies of funders build on and accelerate changes that have been occurring over the last two or three decades in the way that biomedical science is carried out and scientific data are generated and analysed. In genomics, change has been driven primarily by the need for fundamental sequence information, comparative populations, large numbers of samples, and by the falling costs and increasing capacity of sequencing and computing technologies. Research practice has become increasingly interdisciplinary,7 with the rapid formation of flexible and dynamic research collaborations around the world.8 For example, the use of new methodology such as GWAS requires large numbers of clinically well-characterised samples to be collected from patients; laboratory staff and researchers to manage the genotyping pipeline; bioinformaticians, statisticians and other data analysts to interpret the data; and leadership from principal investigators. In combination, these factors have had a significant effect on the way that research projects are planned, organised and managed, and have encouraged the development of open access policies (BOX 1).


Data Sharing Policies

Open access to data is believed to accelerate advances in science, by making data freely available to all, while also ensuring the expedient use of existing resources that have been funded by the public purse. The first international document to embody this perspective and lay out the principles for open access in the field of genomics was the Bermuda Agreement in 1996, which was followed by the Fort Lauderdale Agreement in 2003. These documents together set out the key principles which now dominate thinking and practice regarding open access to genome sequence data in North America and the UK.

The key idea being promoted in the Bermuda Agreement is that the pre-publication genome sequence ‘should be freely available and in the public domain in order to encourage research and development and to maximise its benefit to society.’ The Fort Lauderdale Agreement took this further by setting out a plan of ‘tripartite responsibility’ for sequence producers, users and funders for the establishment of ‘community resources’ to achieve rapid and open data release. This agreement stated that ‘community resource data sets benefit the users enormously, giving them the opportunity to analyse the data without the need to generate it first. The data sets are, in general, much larger, richer and of higher quality than individual laboratories could normally generate.’ Such datasets have been presented as the ‘drivers of progress in biomedical research’ and therefore they should be ‘made immediately available for free and unrestricted use by the scientific community to engage in the full range of opportunities for creative science.’

The open access principles underlying these developments have since been applied by national funding bodies beyond projects which generate sequence data to other areas of biomedical research. Examples of such policies are those of the National Institute of Health (NIH 2003), Genome Canada (2005), and the UK Medical Research Council (MRC 2006). All of these organisations now make data sharing a requirement of funding in genomics. These policies have created a climate in which data sharing has become the default, and applicants must demonstrate why their data should be exempt from the requirement that it should be deposited for use by other scientists.

Hypothesis-led projects

In the case of new hypothesis-led projects, researchers are required to provide, in their funding proposal, a plan for how data and results will be shared. The specific aim of data sharing policies is to ensure maximum availability of data. Arguments can be made for excluding access to the data by some researchers on the basis of the sensitivity of the data, or the potential to identify or stigmatise individuals or groups. While newly funded projects can be planned and developed in accordance with data sharing policies, greater challenges arise, as is indeed happening, when such policies are applied retrospectively to completed projects or to ongoing longitudinal projects.

Large-scale, data-generation projects

Advances in sequencing and computing technologies have also enabled the scientific community to embark on a new type of scientific effort, specifically large-scale, data-generation projects. Such projects generate data and create management infrastructures, or platforms, which can support simultaneous access to a dataset by multiple researchers. Secondary users of the data are far removed from the researchers who carried out the collection of the samples and data, as well as from the research participants. In such projects, research participants are informed that the analysis of their sequence will be freely available on the web. These projects have enormous benefits for the entire scientific community, as they have accelerated the creation of new knowledge and provided a blueprint for data sharing (BOX 2).


Data Generating Projects and Access Criteria

Open access policies

Several data-generating projects provide free access to data online. For example, the Human Genome Project (1990-2003) aimed to sequence the 3 billion base pairs in the human genome and to identify all 20,000-25,000 genes. The HapMap Project (2002-2005) identified chromosome regions with sets of strongly associated SNPs, the haplotypes in those regions, and the SNPs that tag them. The 1000 Genomes Project (which began in 2007) will develop a map of biomedically relevant DNA variations at unprecedented resolution.

Each of these projects has relied upon the co-operation of funders and researchers from many disciplines, and has drawn on considerable resources, expertise and time. As none of these projects provides any link to phenotypic information, access to the data is freely available through the internet, regardless of intended use of the data or identity of the user.

Restricted access policies

By contrast, projects that generate, combine and archive different kinds of data, such as dbGaP, the Genetic Association Information Network (GAIN), and the Wellcome Trust Case Control Consortium (WTCCC), have developed data release policies to control access. Some data are placed on the web, but researchers must establish their credentials before they are allowed access to information that could potentially identify research participants.

The dbGaP is a repository of four types of data: study documentation; phenotypic data; genetic data (including study subjects' individual genotypes); and statistical results, including some association and linkage analyses. dbGaP provides two levels of access - open and controlled - to allow broad release of non-sensitive data, while providing oversight and investigator accountability for data sets involving personal health information. The benefit of dbGAP is that it provides a controlled archiving system for research data.

GAIN (2006-2008) completed an ambitious program to genotype existing research studies in six major common diseases, and to combine results with clinical data to create a significant new research resource. The resulting data are being deposited in a database within the National Library of Medicine at NIH, funded by GAIN, for the broad use of the research community. Originators of the initial studies received additional grants to make their own analyses. Access is controlled by an NIH data access committee.

The WTCCC (2007-) is a collaboration of 24 geneticists based in the UK that is analysing thousands of DNA samples from patients to identify common genetic variations for different diseases. Aggregated data are placed on the internet, but access to the more detailed genotypic and phenotypic data is obtained only through the principal investigator, who can also decide on further collaboration.

The primary goal of all these initiatives is to make data as widely available as possible to further scientific progress. However, decisions about access are centralised and are no longer controlled by the research team who collected the data; instead researchers must conform to specific deposition and access requirements, which in turn affects the way in which research is conducted.

The changing landscape of data sharing

The data sharing policies of funders may crystallise and encourage existing trends in scientific practice. In the past, data sharing has primarily been done with known colleagues, based on mutual respect, trust and a common interest. The conditions of access would be negotiated on an individual basis and would vary according to particular circumstances. Funders now require that data sharing be considered in every newly funded research project, unless there are justifiable reasons why this should not be so. With these policies, the question for many researchers has become how to share data, whereas previously it was whether data should be shared at all. This creates a number of challenges for several areas of scientific practice. We begin by discussing how best to provide rewards and incentives for the researchers who have been involved in data generation.

Acknowledging individual contributions

In the past, a dataset would have been used primarily by the researchers who had created it, and would provide the basis for many publications. There would have been a direct relationship between the creation of the data and control over usage and the publication of results. However, with data sharing policies, the fact that particular researchers have created a dataset no longer gives them an enduring priority or control over its use and resulting publications. The challenge then is how to reward and acknowledge the production of a dataset.

Proper recognition for authors and contributors

The traditional form of acknowledgement is through a publication, which is also a key way of ensuring career advancement. Guidelines for many journals require that data production should be acknowledged, but how this is done is largely left up to individuals following the norms that exist within different disciplines. One solution has been to publish articles with large numbers of authors,9 recognising the involvement of many researchers and data producers in large collaborations. A difficulty arises when the number of authors becomes excessive, as authorship is more a reflection of contribution to a project rather than to a publication. The practice adopted by some publications is to describe the contributions of individual authors, although this policy is difficult to extend to large numbers of authors. An alternative means might be to making a distinction between a ‘contributor’ — who has provided the dataset— and an ‘author’ — who has worked on the analysis or result.

Means of recognition other than traditional authorship have also been proposed.10 In one approach, the dataset itself would be recognised in the publication according to an established system. This would acknowledge use of the dataset and indirectly reward the contributions of those who have been instrumental in establishing the resource, without needing to cite each contributor to the generation of the dataset. Recognition in a publication is essential but data generation needs to be established as an activity worthy of recognition in its own right, which relies on specialist skills. Therefore, it is important that the efforts of data generators are appreciated by the scientific community, and the establishment of a resource for other researchers is considered as a valuable output by institutions. In addition, there must be indices which can also be included in national assessment schemes, such as the Research Assessment Exercise (RAE) in the UK, which ranks institutions according to their research excellence.

Promoting data sharing

Although publications and formal recognition are important, incentives to share data also need to be built into the research process. One solution developed by the Genetic Association Information Network (GAIN) (see BOX 2), is to give the producers of the data a six month publishing lead on their competitors, even though the data are available to all bona fide researchers during this time. The researchers that generate the data are given the opportunity by funders to develop a dataset using new GWA technology. However, this incentive can also place enormous pressure on the research team who are generating data as well as attempting to analyse and publish results within a short time span. Working constantly against rapidly impending deadlines is not in the long run a productive climate, and an extension of such publishing lead times should be considered. Such incentives require careful thought, as they are having a significant effect on the way that science is being conducted, both in terms of teamwork and in the speed of data generation.

Novel ways of acknowledging contributions to the generation of data are required, which are fair and transparent, lest researchers obstruct data sharing. Genomic data is only useful for subsequent analyses if it is accompanied by good metadata that describes, for instance, sample collection procedures, clinical definitions of the cases, and demographic data. In practice, therefore, scientists can retain some measure of control over access,11 for example, by claiming that part of the dataset is not ready to be shared. This would make it difficult for other researchers to do meaningful analyses.12 This is contrary to the principles of data sharing and very hard to guard against. However, it would be inappropriate and cumbersome to develop punitive oversight mechanisms to ensure this does not happen. Instead, ‘carrots’ rather ‘sticks’ need to be used to encourage those that create metadata to share with others further downstream in the scientific process.

Such incentives need to replicate the climate of trust and reciprocity that accompanies traditional and more informal data sharing. No one wants to be part of a system where they feel someone else can take advantage of their unsung contributions. One way forward is to have an open and honest debate within the scientific community on how and why data sharing, both formal and informal, works or does not work. This debate is necessary to articulate the norms required in specific situations to determine a fair and equitable way to share data but also acknowledge individual contributions. This is not a matter of more regulation and guidelines, but developing norms that become an intrinsic part of a new scientific culture, in which people can trust each other because the rules and obligations are known at the outset.

Responsibilities towards study participants

The original context in which the samples and data are collected is associated with expectations and relationships that are understood both by researchers and participants.13 Researchers may feel a strong sense of responsibility for ‘their’ samples and feel a legal and moral responsibility for research participants that often extends beyond the original terms of consent. This responsibility may not be felt by secondary researchers who have no connection with the research participants, and see themselves as only dealing with data. While secondary researchers have an obligation to use data in a scientifically sound, ethical and lawful manner, these obligations are not the same as the researchers enrolling patients in a study. Informed consent forms, which try to be succinct, may not embody all of the expectations that are associated with enrolment in a study and an on-going clinical relationship, and may leave room for differing interpretations of the scope of consent.

In data-sharing policies, researchers are given the opportunity to justify why raw data should not be shared. Given their knowledge about the types of uses that may be made of data, based on the original consent, researchers are in a good position not only to decide on appropriate uses, but also to protect against possible misuse. In particular, when samples are collected and analysed as an extension to ongoing epidemiological work, cohort studies, or disease-specific work where the relationship develops in a clinical setting, the obligation to share genomic data may be perceived as an imposition on the relationships that have been built up between researchers and participants.

The challenge for funders is to ensure that this sense of stewardship is respected, by ensuring that new systems for sharing data acknowledge these perceived responsibilities. There is a danger that data sharing policies may be experienced as being punitive, or that those who feel uncertain about sharing may be characterised as obstructive and short-sighted. However, reluctance to share may have sound justifications. These claims should not be automatically dismissed as researchers being territorial. Such concerns cannot be ignored, as they can have practical as well as ethical implications in a project, where the trust and support of participants is vital. In addition, researchers who are perceived to be uncooperative could be excluded from key areas of activity, such as developing strategic policies and being involved in peer-review. Funding bodies must be prepared to consider the claims of those for whom data sharing, because of the nature of their research and situation, may create difficulties. At the moment, requests for exemption from data sharing are judged by funders but it may be better for this assessment to be made by bodies which are independent from considerations about future funding for the applicants.

Maintaining Public Trust

The mechanisms that have traditionally been used to protect research participants are informed consent and the anonymisation of datasets. However, the sharing of data from genomic studies tests the effectiveness of these standard mechanisms of privacy protection.

Anonymisation of data

The digital revolution, which has allowed many types of data to be shared both with and without consent, is rapidly changing the landscape of privacy protection. Procedures for controlling disclosure, such as coding each study subject or aggregating the information, can be employed to protect the identity of data subjects. However, these policies may lessen the scientific utility of the data, as fine detail and nuances can be lost in the effort to protect privacy.14 Furthermore, as DNA is a unique identifier, it is impossible to completely anonymise a sample, and relatively small numbers of SNPs can be used to identify individuals.15 The recent decisions by the Wellcome Trust and the NIH to remove SNP data from publicly accessible databases, following the paper by Homer and others,16 illustrate the problems of protecting participants' privacy interests whilst using GWAS methodology. Homer and his colleagues established by a statistical analysis that an individual could be identified in aggregate data, as genome-wide scans provide such a wide range of unique data points.

Informed consent

The process of obtaining informed consent is one way for research participants to have some control over how their information is used: however, this procedure is problematic when it is applied in a data sharing context.

First, it is difficult to achieve the level of understanding required for truly informed consent,17 -19 especially where data sharing in genomics is concerned: participants have a variable understanding of whether their sequence data will be shared, and with whom.20 Second, it is difficult to provide information about all the potential users of shared data, without a constantly updated system to inform participants. Many long-term studies, such as the Avon Longitudinal Study (ALSPAC) have approached this problem through web sites for participants that enhance an understanding of the science. Greater patient involvement in the decision-making of biobanks has also been proposed in order to compensate for this deficit. 21 (BOX 3)



Models such as broad consent have been proposed as a solution to some of the ethical challenges of data sharing. In broad consent, an individual gives consent to widely specified research, which allows for many future uses of tissue and data rather than just the one (or more) use(s) specified by known researchers. Once individuals give consent, they are not re-contacted concerning new uses. In projects where there is uncertainty about the scope of the consent, authorisation for the use of coded samples and data may be given by a research ethics committee.

However, there is concern about whether this practical solution to the issue of informed consent is compliant with data protection principles,22 which require that the individual should know how, and by whom, their data are being processed.

There is also concern that a broad consent undermines one of the fundamental principles of medical research, that of individual autonomy and the right of individuals to decide the nature of their involvement in medical research.23 There are differing conceptions of autonomy, however; in some views, individual autonomy requires decisions to be based upon full information; according to others, full information is not required for autonomous consent as long as individuals understand the broad nature of what is proposed and understand that they do not have all the details of what is involved. However, this latter situation demands a greater level of trust in the individuals and institutions concerned.

Data sharing challenges existing mechanisms for privacy protection. Once data have been released into the public domain, participants and researchers have little or no control over their usage, or the possibility that they may be linked to other data sets. Research participants can exercise only consent or withdrawal: however, it is difficult for participants to control how their genomic data will be shared – typically they are required to consent to all data sharing between researchers or none. In addition, there are real doubts whether an individual's request for withdrawal can be meaningful, due to the complexity of retracting data through different datasets. In this new context of global data-sharing, better methods of informing participants about the use of their personal information for different research purposes need to be developed.

Oversight of Access

Data sharing raises new dilemmas for the oversight of research and for the bodies that have been entrusted to ensure that research is well governed. Traditionally, approval for research is obtained from a research ethics committee by a particular individual or research group. This committee holds the principal applicant responsible for monitoring the use of samples and data; however, when samples are transported across national borders, and when data are analysed by people who bear no relation to the original research project or participants, it is virtually impossible to continue to hold the original collector responsible in the same way. Therefore it is difficult for research ethics committees to exert their original mandate to ensure the ethical conduct of research.

It has recently been suggested that ethics committees could have an important function in monitoring the particular ethical aspects of GWA studies. Although such responsibilities have mainly focused on the consent process24 and on approving the re-use of samples,25 some have also considered the potential contribution that ethics committees could make to regulating data access.26,27 However, ethics committees are already facing increasing challenges in reviewing complex research proposals28, and it is therefore not clear how already over-burdened committees could take on the role of monitoring and approving data access –a task that requires significant insight into the techniques used to produce and analyse data in genomics.

Instead of relying on research ethics committees, some data sharing initiatives have established specialist advisory bodies, or ‘data access committees’, to determine who should have access to data and on what grounds. These decisions are not applied uniformly, and the criteria for decision-making are still in their infancy. However, many of the publicly available criteria for determining access seem to involve establishing whether a scientist is a ‘bona fide researcher’, rather than considering whether access will have an effect on research participants. The criteria to qualify as a ‘bona fide researcher’ are still being developed but could include type of research, or are simply based on verifying an individual's credentials such as institutional affiliation. In the case of the NIH dbGaP users are granted a Data Use Certificate which allow access to several datasets. While this generic authorisation may be efficient, there is the possibility that the privacy interests of research participants may be overlooked when a number of datasets are linked together (BOX 4).


Global Data Sharing

Organisations such as P3G, the Biobanking and Biomolecular Resources Research Infrastructure (BBMRI), and the Organization for Economic Cooperation and Development (OECD) have started the legal analysis that is required to develop mechanisms that promote global data-sharing while at the same time ensuring that research is carried out ethically and according to accepted standards. Ideally, the new framework would relieve researchers from having to seek approval from multiple data access committees, but this direction is still being debated. One possibility is the development of a system where an approval for access is given by one international body for a number of similar projects rather than having independent access committees for each project. This could develop uniformity in decision-making and create a clear and transparent set of criteria for deciding questions of access for all researchers. The disadvantage is that it removes decision-making from the local level to a body that is removed from the context in which the dataset has been established.

One of the problems of such a proposition is that while international agreements can help to set broad standards, all countries have their own systems of law. This means the flow of data and samples through a number of countries will be subject to many different legal regimes, and to different sets of guidelines and standards. The concept of an international body to oversee data sharing is good in theory, but in reality it would probably add another layer of bureaucracy for researchers, as they would be forced to comply with the international layer of approval, as well as comply with national regulations.


Data-sharing policies have been in place in the USA for the past six years and in the UK for approximately three years, and their full effect on scientific practice is just starting to be understood. The challenge for individual researchers is to simultaneously fulfil the requirements of funding bodies, honour their obligations to study participants, and protect their own interests and careers. The challenge for funders is to ensure that public trust is maintained and that data sharing policies improve the transfer of research results and knowledge. To ensure that this framework is sustainable, good working relations are required between funders, collaborating groups of scientists and the many thousands of recruits that will be needed in the future.

Establishing and maintaining global public goodwill and trust is an ongoing task to ensure the future of sound, and hence ethical, scientific research. Meeting such challenges is necessary to ensure that data sharing practices and policies continue to produce the harvest of tangible benefits currently enjoyed by some of the scientific community. However, further research and thought is required to address some of the challenges to scientific practice that data sharing creates.


We would like to thank the anonymous reviewers for their insightful comments and suggestions. JK, CH, NH and JdV are funded by the Wellcome Trust, under grant codes WT 081407/Z/06/Z, WT 076070/Z/04/Z,WT 077869/Z/05/Z and WT 083326/Z/07/Z. PB is funded by EU FP6, Procardis Project number 037273.



Jane Kaye

Jane Kaye is a lawyer by training (DPhil Oxon; Admission to Practice (ACT Supreme Court); LLB (Melb.) BA (ANU)) and is the Wellcome Trust Research Fellow in Law at the Ethox Centre in Oxford, UK. Within the Ethox Centre, Dr Kaye is responsible for directing the Law, Health and Emerging Technologies research programme and she is also Director of the Oxford Bioethics Network (OxBioNet) of the University of Oxford. She is a member of the Faculty of Law at the University of Oxford and is Co-Chair of the P3G International Working Group on Ethics, Governance and Public Engagement.


Catherine Heeney

Catherine Heeney's first degree was in Philosophy from Dundee University, UK, after which she obtained a Masters in Philosophy from Liverpool University in 1999. She then undertook an ESRC (Economic and Social Research Council) PhD Case Studentship in the Centre for Census and Survey Research at the University of Manchester, where she obtained her doctorate in 2004. In 2003 Catherine became a Research Fellow on the ESRC funded Data-Sharing and Privacy project based at Edinburgh University. In 2005 she took up a post at ESRC Genomics Policy and Research Forum at Edinburgh University, before moving to the Ethox Centre to work on the Governing Genetic Databases project (2006-9).


Naomi Hawkins

Naomi Hawkins obtained her BSc (Biomedical Science) and her LLB from the University of Queensland, Australia, in 2002. Following a period of legal practice in Australia clerking for a Supreme Court Judge, and working in a large commercial law firm, she completed her BCL (a Masters degree in law) at the University of Oxford, UK, in 2005. She is currently completing her doctorate in law at the University of Oxford with Dr Jane Kaye and Prof. Colin Tapper, focusing on the impact of human gene patents on the development and use of genetic diagnostic tests.


Jantina de Vries

Jantina de Vries holds a BSc and MSc in Rural Development Sociology from Wageningen University in the Netherlands, alongside qualifications in Biology at the University of Groningen. Jantina also obtained a research master degree in Social and Political Sciences at the European University Institute in Florence, Italy. She currently holds a Wellcome Trust Research Studentship (2008-2011), and is supervised by Professor Michael Parker and Professor Ray Fitzpatrick. Jantina is part of the MalariaGEN Ethics Team. This project is led by Prof. Dominic Kwiatkowski and is based in the Wellcome Trust Centre for Human Genetics in Oxford.


Paula Boddington

Paula Boddington started her career teaching philosophy at Bristol University, having completed her BA in philosophy and psychology at Keele University and her doctorate in philosophy at Corpus Christi College, Oxford, UK. She then moved to the Australian National University, where she taught and researched applied ethics. She subsequently developed interests in medical ethics and specifically in genetics at the Centre for Economic and Social Aspects of Genomics (CESAGen), Cardiff, also gaining a masters degree in medical law from Cardiff University. She is currently providing ethics support for the Procardis programme, a genome-wide strategy to identify susceptibility loci in precocious coronary artery disease.


1. The Digital Archiving Consultancy (DAC) The Bioinformatics Research Centre, University of Glasgow (BRC) The National e-Science Centre (NeSC) Large-scale Data Sharing in the Life Sciences: Data Standards, Incentives, Barriers and Funding Models. Aug, 2005. The “Joint Data Standards Study”. (2005) (Accessed 05/07/08)
2. McCarthy MI, Abecasis GR, Cardon LR, Goldstein DB, Little J, Ioannidis JPA, Hirschhorn JN. Genome-wide association studies for complex traits: consensus, uncertainty and challenges. Nature Rev. Genet. 2008;9:356–369. [PubMed]
3. Gibbons SMC, Kaye J, Smart A, Heeney C, Parker M. Governing genetic databases: Challenges facing research regulation and practice. J. Law Soc. 2007;34:163–89.
4. Couzin J. Genetic privacy. Whole-genome data not anonymous, challenging assumptions. Science. 2008;321:1278. [PubMed]
5. DNA databases shut after identities compromised. Nature. 2008;455:13.
6. Kaiser J. House weighs proposal to block mandatory ‘Open Access’ Science. 2008;321:1621. [PubMed]
7. Ziman J. Real Science: What It Is, and What It Means. Cambridge: Cambridge University Press; 2000.
8. Gibbons M, Limoges C, Nowotny H, Schwartzmann SSP, Trow M. The New Production of Knowledge. The Dynamics of Science and Research in Contemporary Societies. London: Sage Publications; 1994.
9. Wellcome Trust Case Control Consortium Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature. 2007;447:661–678. [PMC free article] [PubMed]
10. Cambon-Thomsen A. Assessing the impact of biobanks. Nature Genet. 2003;34:25–26. [PubMed]
11. Campbell EG, Clarridge BR, Gokhale M, Birenbaum L, Hilgartner S, Holtzman NA, Blumenthal D. Data withholding in academic genetics: evidence from a national survey. JAMA. 2002;287:473–80. [PubMed]
12. Foster MW, Sharp RR. Share and share alike: deciding how to distribute the scientific and social benefits of genomic data. Nature Rev. Genet. 2007;8:633–639. [PubMed]
13. Nissenbaum H. Protecting privacy in an information age: The problem of privacy in public. Law Philos. 1998;17:559–596.
14. Karp DR, et al. Ethical and practical issues associated with aggregating databases. PLoS Med. 2008;5:e190. doi:10.1371/journal.pmed.0050190. [PMC free article] [PubMed]
15. McGuire AL, Gibbs RA. Genetics. no longer de-identified. Science. 2006;312:370–371. [PubMed]
16. Homer N, et al. Resolving individuals contributing trace amounts of DNA to highly complex mixtures using high-density SNP genotyping microarrays. PLoS Genet. 2008;4:e1000167. [PMC free article] [PubMed]
17. Lunshof JE, Chadwick R, Vorhaus DB, Church GM. From genetic privacy to open consent. Nature Rev. Genet. 2008;9:406–411. [PubMed]
18. McGuire AL, Caulfield T, Cho MK. Research ethics and the challenge of whole-genome sequencing. Nature Rev. Genet. 2008;9:152–156. [PMC free article] [PubMed]
19. Caulfield T, et al. Research ethics recommendations for whole-genome research: consensus statement. PLoS Biol. 2008;6:e73. [PMC free article] [PubMed]
20. McGuire AL, Hamilton JA, Lunstroth RJD, McCulloch LB, Goldman A. DNA data sharing: research participants' perspectives. Genet. Med. 2008;10:46–53. [PMC free article] [PubMed]
21. Winickoff DE. Partnership in U.K. Biobank: A third way for genomic property? J. Law Med. Ethics. 2007;35:440–456. [PubMed]
22. Zika E, et al. Sample and data use and protection in biobanking in Europe: legal issues. Pharmocogenomics. 2008;9:773–781. [PubMed]
23. Caulfield T. Biobanks and blanket consent: The proper place of the public perception and public good rationales. King's Law J. 2007;18:209–226.
24. Mascalzoni D, Hicks A, Pramstaller P, Wjst M. Informed consent in the genomics era. PLoS Medicine. 2008;5:e192. [PMC free article] [PubMed]
25. Muula AS, Mfutso-Bengo JM. Responsibilities and obligations of using human research specimens transported across national boundaries. J. Med. Ethics. 2007;33:35–38. [PMC free article] [PubMed]
26. Upshur R, Lavery J, Tindana P. Taking tissue seriously means taking communities seriously. BMC Med. Ethics. 2007;8:11. [PMC free article] [PubMed]
27. Langat SK. Reuse of samples: ethical issues encountered by two institutional ethics review committees in Kenya. Bioethics. 2005;19:537–549. [PubMed]
28. Hoedemaekers R, Gordijn B, Hekster Y, Van Agt F. The complexities of ethical evaluation of genomics research. HEC Forum. 2006;18:18–36. [PubMed]

Further Reading Online