Ethics, Legal & Social Issues (ELSI) in Genomics

Short version. ELSI — ethical, legal, and social implications — entered genetics with the Human Genome Project. The NHGRI ELSI Research Program was established in 1990 alongside the HGP, with up to 5% of the project budget reserved for studying its consequences; it remains the world's largest funder of bioethics research. The contemporary ELSI conversation covers consent in research and clinical settings, the patchwork of statutory protections (GINA 2008 in the US; GDPR Article 9 in the EU and UK), the duty-to-warn debate (the ABC v St George's litigation in the UK), forensic familial DNA searching (Bieber, Brenner & Lazer 2006, Science 312:1315; the Golden State Killer case in 2018), regulation of direct-to-consumer testing (the FDA's 2013 warning letter to 23andMe), equity in genomic data (Sirugo, Williams & Tishkoff 2019, Cell 177:26), newborn-screening expansion (Genomics England's Newborn Genomes Programme), and polygenic embryo selection (Karavani et al. 2019, Cell 179:1424; Turley et al. 2021, NEJM 385:78).

The ELSI Research Program of the NHGRI

When the US National Human Genome Research Institute (then the Office of Genome Research) launched the Human Genome Project in 1990, it allocated 3–5% of the project's budget to research on the ethical, legal, and social implications of human genetic research. The intention, articulated by James Watson at the launch and elaborated by his successors, was to anticipate the social consequences of mapping the human genome before they crystallised in policy or jurisprudence. The reservation was unusual: most large research programmes do not fund critical study of themselves.

Three decades on, ELSI is an established academic field with dedicated journals (the American Journal of Bioethics, Genetics in Medicine's ELSI articles, Bioethics, Hastings Center Report), training pathways, and a corpus of empirical and conceptual literature on consent, privacy, public engagement, equity, return of results, biobank governance, and the social impact of genomic technologies. The NHGRI ELSI Research Program remains the largest single funder of bioethics in the world.

Informed consent in research and the clinical setting

The modern era of consent in genomics rests on three foundations: the international research-ethics tradition descending from the Nuremberg Code (1947), the Declaration of Helsinki (1964 and subsequent revisions), and the Belmont Report (1979); the contemporary regulatory implementation in the US Common Rule (45 CFR 46, revised 2018) and equivalents elsewhere; and statutory protections specific to genetic information (GINA in the US; the special-category provisions of GDPR in the EU and UK). The literature has converged on a series of distinctive features of consent in genomics: the implications for biological relatives who have not consented; the long persistence of identifiability in genomic data even after notional de-identification (Gymrek et al. 2013 demonstrated surname inference from Y-chromosome haplotypes); the changing meaning of a result over time as the literature on a variant evolves; and the difficulty of full consent for unspecified future research uses in biobank settings.

Tiered, dynamic, and broad-consent models have been proposed in response. Each is taught alongside the empirical literature on what participants actually understand and want; the Genomics England 100,000 Genomes Project and the US All of Us programme have generated substantial empirical material on consent in large-scale national programmes.

GINA 2008 (US)

The Genetic Information Nondiscrimination Act of 2008 (Pub.L. 110-233) prohibits discrimination on the basis of genetic information in two domains: health insurance (Title I) and employment (Title II). GINA prohibits health insurers from using genetic information to set premiums or deny coverage; prohibits employers from requesting genetic information for employment decisions; and limits the disclosure of genetic information held by employers and insurers. The protections were a precondition for the public expansion of clinical genetic testing in the US, but are deliberately limited: GINA does not apply to life insurance, long-term care insurance, or disability insurance, where genetic-test results may legally be used in underwriting in many states. The literature documents continued patient anxiety about life insurance as a barrier to clinical and research testing despite GINA's existence in the health and employment domains.

GDPR Article 9 and the UK ICO

Under the EU General Data Protection Regulation (GDPR) Article 9 and the equivalent UK provisions retained after Brexit (UK GDPR), genetic data is treated as a "special category" of personal data, with stricter conditions for lawful processing than ordinary personal data. Processing requires the data subject's explicit consent or one of a defined set of public-interest exceptions (medical, scientific research, public-health). The UK Information Commissioner's Office publishes guidance on how Article 9 applies in research and clinical contexts. The literature on GDPR-and-genomics covers the lawful basis for biobank research, the practical operation of the public-interest research exception, the cross-border-transfer rules, and the right of access (Article 15) and erasure (Article 17) and how those rights interact with the public-good of long-term genetic-data linkage.

Patient-facing pedigree-drawing and family-history-collection software inherits GDPR obligations: the data being collected is special-category genetic data both about the patient and about their biological relatives. Evagene's privacy posture is described in the site privacy policy; in summary, the platform is a research, education, and family-history-documentation tool, the patient retains control of the file and chooses where it goes, and clinical-grade computation that the platform itself does not provide (such as BOADICEA via the CanRisk file bridge) is run off-platform by the user.

Forensic familial DNA searching

The use of DNA databases to identify suspects through their relatives — "familial DNA searching" or, more recently, "forensic genetic genealogy" using consumer-genealogy databases — has been one of the most ELSI-charged developments of the past two decades. The technique was first set out as a structured public-policy possibility in Bieber, Brenner & Lazer 2006 (Science 312:1315); the UK National DNA Database had begun limited familial searching shortly before that paper.

The 2018 identification of Joseph James DeAngelo as the Golden State Killer through a DNA profile uploaded to the genealogy database GEDmatch — followed by the construction of a family tree from partial-match relatives, and the eventual identification of DeAngelo by triangulation — brought the technique into mainstream attention and triggered a substantial bioethics, legal, and policy literature. The ethical issues raised include: the consent of the relatives who uploaded their data with no expectation of forensic use; the risk of false identification given the demographic biases of the consumer-genealogy databases; the racial-equity question (the underlying databases are predominantly European-ancestry, with consequences for who can be identified); and the constitutional and data-protection law as it applies to a third-party-held database. The literature on forensic genetic genealogy has continued to grow rapidly through the 2020s.

The duty-to-warn debate: ABC v St George's

Whether and how a clinician's duty extends from the proband to the proband's at-risk relatives was the subject of a sustained UK litigation, ABC v St George's Healthcare NHS Trust and others. The claimant, ABC, alleged that her father's clinicians had a duty to inform her of the father's diagnosis of Huntington's disease in time for her to consider termination of an established pregnancy; the father had refused permission for disclosure. The case made law over a decade: the Court of Appeal held in 2017 that a duty might be arguable, the High Court ruled in 2020 that on the specific facts the standard of care had not been breached, while accepting that the duty in principle could exist (ABC v St George's Healthcare NHS Trust [2020] EWHC 455 (QB)). The judgment is the leading UK authority and is taught alongside the American Society of Human Genetics 1998 statement (Am J Hum Genet 62:474), which set out the structured ethical position that a clinician's primary duty is to the proband but in narrowly-defined circumstances disclosure to a relative may be ethically defensible.

23andMe and the regulation of direct-to-consumer testing

Direct-to-consumer (DTC) genetic testing entered mainstream attention through 23andMe's launch in 2006 and the related companies that followed. The FDA's November 2013 warning letter to 23andMe ordered the company to stop marketing its Personal Genome Service for health-related uses on the grounds that it was an unapproved medical device. After a multi-year process the FDA in 2017 authorised 23andMe to market a small set of carrier-status reports and in 2018 a small set of pharmacogenomic and disease-predisposition reports under specific conditions; the regulatory pathway has continued to evolve.

The DTC story is taught as a case study in how a fast-moving consumer technology interacts with a slow-moving regulatory apparatus; in the implications of marketing a clinical-purpose product directly to consumers without clinical interpretation; and in the data-protection consequences of consumer genomic databases (the 2023 23andMe data breach exposed the genetic data of millions of users). The literature includes empirical studies of how consumers act on DTC results, the rate of false reassurance and false alarm, and the burden DTC results place on clinical-genetics services downstream.

Secondary and incidental findings (revisited)

The secondary-findings question described on the diagnostics and counselling page is also an ELSI question. The ACMG SF list (Kalia et al. 2017, Genet Med 19:249; updated to v3.2 by Miller et al. 2023 and to v3.3 by Miller et al. 2025, Genet Med 27:101391, adding ABCD1, CYP27A1, and PLN) is justified on the grounds that the findings are medically actionable and the burden-versus-benefit calculation falls on the side of return; the bioethics literature has interrogated each of those premises (what counts as actionable; what counts as benefit; how the right not to know is preserved) and continues to do so.

Equity and diversity in genomic data

Sirugo, Williams & Tishkoff 2019 (Cell 177:26) documented the persistent under-representation of non-European ancestry in genomic data: roughly 78% of GWAS participants at the time of the survey were of European ancestry, despite Europeans accounting for ~16% of the world's population. The under-representation has direct technical consequences: polygenic risk scores derived in European-ancestry cohorts perform poorly when applied to African, East Asian, and other ancestry groups; allele-frequency reference databases such as gnomAD systematically under-call rare variants in under-represented ancestries; clinical variant interpretation is biased toward variation seen in European-ancestry patients.

The response has been a coordinated push toward diversification of genomic resources. Major programmes:

H3Africa (Human Heredity and Health in Africa) — a pan-African research consortium, jointly funded by the NIH and the Wellcome Trust, building African-centred genomic infrastructure, biorepositories, and capacity.
All of Us — the US national programme aiming to enrol one million participants with deliberate over-recruitment from communities historically under-represented in biomedical research; the genomic dataset is structured to support diverse-ancestry research.
Our Future Health — the UK's largest health-research programme, planned for ~5 million participants, with a deliberate target for ethnic-minority participation reflective of the UK population.
Population biobanks elsewhere — FinnGen (Finland), BioBank Japan, Singapore's PRECISE programme, Saudi Arabia's national genomic project, and many others.

The ELSI literature on equity covers not only the technical equity question (is the science the same?) but also the procedural equity question (are the researched communities partners or subjects?), the benefit-sharing question, and the long-term governance question of who controls what is done with the data.

Newborn-screening expansion

Population newborn screening (Guthrie cards / heel-prick) has been part of public-health practice in many countries since the 1960s, beginning with phenylketonuria. The conditions screened for have expanded steadily since then, governed in most jurisdictions by criteria descended from the Wilson and Jungner WHO 1968 principles: well-defined disease, latent stage with effective treatment, available test, acceptability to the population, cost-benefit balance.

The proposal to extend newborn screening to genome sequencing — rather than the targeted biochemistry of the Guthrie panel — is one of the most active ELSI debates of the 2020s. Genomics England's Newborn Genomes Programme, launched in 2022, is enrolling ~100,000 newborns to explore the feasibility, acceptability, and clinical utility of whole-genome sequencing at birth; comparable programmes are underway in the US (BabySeq, GUARDIAN) and elsewhere. The Wilson and Jungner criteria are being re-examined in this context: what counts as a screenable condition when the technology is broad-spectrum? What is the right model of consent (parental, with the child's own consent at adulthood for adult-onset findings)? What is the appropriate breadth of conditions to report? The empirical literature is in active development.

Polygenic embryo selection

Polygenic risk scores (PRS) summarise the contribution of many common variants to a complex trait. The proposal to apply PRS to embryos generated through IVF, ranking and selecting embryos on predicted polygenic risk for a disease (or, more controversially, predicted polygenic value for a trait such as cognitive ability or height), entered the literature and the consumer-clinic offering in the late 2010s.

The technical limits are quantified in Karavani et al. 2019 (Cell 179:1424), which simulated polygenic embryo selection for height and IQ across realistic IVF cohort sizes and showed that the expected gain (typically 2–3 cm of adult height; ~3 IQ points) is modest, much smaller than parental-mid-parent variation, and dwarfed by environmental and family-level confounders. Turley et al. 2021 (NEJM 385:78) reviewed polygenic embryo selection from clinical, ethical, and policy perspectives and reached comparable conclusions: the technical performance is poor, the ethical concerns are substantial, the regulatory framework is patchwork, and the marketing has generally outrun the evidence.

The ELSI conversation around polygenic embryo selection covers the technical-validity question (the modest predicted gains and the reliance on European-ancestry-derived scores), the ethical questions (eugenic-shadow framing; the parental-autonomy framing; the trait-vs-disease distinction), and the regulatory questions (selection on disease risk and selection on trait value are differently regulated in many jurisdictions; the ESHRE position differs from the more permissive US private-clinic context).

How this connects to teaching practice

ELSI is not an add-on to clinical-genetics training but a constitutive part of it. The Resta 2006 definition of genetic counselling — "the process of helping people understand and adapt to the medical, psychological and familial implications of genetic contributions to disease" — foregrounds the family-and-society dimension that ELSI articulates. The clinical genetic skills page describes the family-history interview that is the entry point; the diagnostics and counselling page describes the test, the variant, and the disclosure conversation; this page describes the framework of consent, equity, regulation, and emerging-frontier debate within which all of that takes place.

Boundary statement

This page is a teaching survey of how ELSI in genomics is discussed in the published literature. It is descriptive of the field, not advice on how to act in any particular case; clinical, legal, and personal decisions in any specific situation sit with the people responsible. Evagene is a research, education, and family-history-documentation tool; it is not a medical device, not clinical decision support, and not a diagnostic, screening, or treatment tool. Where this page links to Evagene calculator pages, those calculators are illustrative implementations of published algorithms intended for research and teaching; clinical-grade computation, where available, is provided off-platform.

Ethics, legal, and social issues (ELSI) in genomics