National DNA Database

A register of profiles compiled from crime scenes and from people — and the question of who stays in it, and why, has never stopped being contested.

What the database actually is

A national DNA database is not a store of biological material. It is an index of profiles — lists of numbers at named genetic sites — held alongside an identifier that links each profile to a person or a sample. In England and Wales, that index is the National DNA Database (NDNAD), administered by the Home Office and operational since April 1995. The United States equivalent, CODIS (Combined DNA Index System), went live the same decade under FBI Laboratory management, operating as a tiered structure of local, state and national indices. Both work on the same principle: a crime-scene profile is searched against the collection of person profiles, and a candidate match is flagged for human review.

What the database contains varies by jurisdiction. NDNAD holds profiles obtained from people arrested for a recordable offence in England and Wales, profiles taken from crime scenes (labelled subject unknowns), and profiles of volunteers and elimination samples submitted by laboratory staff. The National Institute of Standards and Technology maintains the reference data for the core loci that profiles must cover — currently twenty autosomal STR sites in the US system, expanded from thirteen in 2017, with overlapping but not identical sets in use in Europe. ENFSI coordinates agreement across European national systems so that cross-border searches through the Prüm framework can work at all.

The profile itself is a short tandem repeat (STR) profile, generated by PCR amplification and read on a capillary electrophoresis instrument. At each locus the instrument reports two allele lengths — one from each chromosome — expressed as a number. A profile is therefore a list of paired numbers, perhaps twenty pairs long. Nothing in that list reveals disease susceptibility, physical appearance or ancestry in any direct, interpretable way, though the debate about what STR data could theoretically disclose remains open in the scientific literature. The biological sample from which the profile was derived is, in England and Wales, generally destroyed after profiling, though this has not always been the case.

CROSS-REF

The argument about retention

The most consequential legal contest over retention concerned two people who were never convicted of any offence. S. and Marper v. the United Kingdom, decided by the Grand Chamber of the European Court of Human Rights in December 2008, held that indefinite retention of DNA profiles, cellular samples and fingerprints from people acquitted or never charged was a violation of the right to private life under Article 8 of the European Convention on Human Rights ↗. The UK at the time retained profiles from arrested individuals regardless of outcome — a policy the court found could not be justified in a democratic society.

The judgment forced legislative change. The Protection of Freedoms Act 2012 introduced time limits: profiles of adults convicted of serious offences are retained indefinitely; profiles of adults arrested but not convicted are generally deleted, with exceptions for certain offence categories; profiles of children are subject to shorter periods. Scotland had already operated a more restrictive regime. The practical effect on database size was substantial — hundreds of thousands of profiles were removed in the years following the legislation.

The United States has its own version of this argument, structured differently because CODIS eligibility is set by federal statute and individual state laws rather than a single court ruling. Most states permit retention of profiles from people convicted of qualifying offences; some collect profiles on arrest; the question of juveniles, expungement on acquittal and the scope of qualifying offences varies state by state. The President's Council of Advisors on Science and Technology and the National Academy of Sciences have both reviewed forensic DNA practice, and the adequacy of legal safeguards around database retention has featured in both bodies' published work.

Familial searching — querying the database not for the contributor but for a partial match suggesting a close relative — extends the population effectively captured by the database far beyond its nominal membership. A person who has never been arrested appears in no index, but if a sibling is retained, a familial search can reach them. This implication is documented and discussed in the forensic genetics literature; the legal frameworks governing it differ sharply between jurisdictions, and several US states prohibit or restrict the practice.

What a match does and does not establish

A database hit is not a conviction. It is the identification of a candidate — a profile in the index that matches the crime-scene profile at the tested loci within stated tolerances. That match must then be evaluated: the likelihood ratio expressing how much more probable the evidence is if the person contributed the sample than if a random unrelated individual did, the possibility of transfer or contamination, the question of whether the biological material carries information about when or how it was deposited.

Secondary transfer — DNA moving from a surface to an object without direct contact from the contributor — is documented and reproducible in laboratory conditions. A profile on an item does not place its contributor at a scene; it records that the biological material was there. This distinction matters more when the database is large, because a larger index produces more adventitious hits — matches that are coincidental or the result of contamination rather than presence. The NDNAD's own annual statistics ↗ record the number of profiles loaded, the number of crime-scene-to-person matches generated, and the number of crime-scene-to-crime-scene links — but not how many of those matches led to a conviction, or a correct one.

Both work on the same principle: a crime-scene profile is searched against the collection of person profiles, and a candidate match is flagged for human review.

Elimination databases — collections of profiles from laboratory staff, police officers and crime-scene investigators — exist precisely because contamination is real and documented. A match that turns out to be a technician's profile is not intelligence; it is a failure mode. SWGDAM's interpretation guidelines and the Forensic Science Regulator's statutory code both address the controls required to catch such failures before they propagate into casework.

The Innocence Project's post-conviction review work has demonstrated that database evidence, like any forensic evidence, can be wrongly used: the question is not whether DNA is reliable in principle but whether a specific result was generated and interpreted correctly, and whether the biological connection it records was correctly understood as evidence of the act charged. The technique's power to exonerate — the first use in a criminal investigation, in Leicestershire in the 1980s, cleared a man before it implicated another — is part of the same record as its capacity to mislead when misunderstood.

A retained profile index is a technical infrastructure and a set of policy choices simultaneously. The infrastructure is well understood. The policy choices — who goes in, who comes out, what a hit authorises — are still being argued, and correctly so.