NIST core loci

The sites everybody reads are a choice — and the institution that maintains the reference data behind them is the National Institute of Standards and Technology.

Why a shared list matters

A DNA profile is only searchable if the laboratories producing it have read the same chromosomal addresses. In the early years of forensic DNA, different jurisdictions ran different sets of short tandem repeat loci, which made cross-laboratory comparison fragile and database searches unreliable. The solution was a mandated core — a fixed list of STR sites that every accredited laboratory in a given system must include. In the United States, the FBI Laboratory defined the original CODIS core in 1997: thirteen loci, chosen because they provided high discrimination power, had well-characterised population data, and could be multiplexed efficiently on the instruments then available.

Those thirteen held for nearly two decades. Then, in 2017, the CODIS core expanded from thirteen to twenty loci ↗, driven partly by the need for better discrimination in a database that had grown to tens of millions of profiles, and partly by the desire to align more closely with European standards. That alignment matters: the Interpol gateway depends on laboratories sharing at least a subset of sites, and expanding the American core closed some of the gap.

What NIST actually provides

Deciding which loci to include is only the first step. For a result at any locus to be meaningful, the laboratory needs to know how common each allele is in the population — and that requires reference data from large, well-documented surveys. The National Institute of Standards and Technology maintains the reference databases and Standard Reference Materials that underpin every American forensic profile. Its STRBase resource has for years catalogued population data, kit performance characteristics and allele frequencies across dozens of population groups, giving laboratories the denominators they need to calculate a likelihood ratio and report how much more probable a profile is given the prosecution hypothesis than the defence hypothesis.

NIST also produces the physical reference materials — characterised DNA samples with known genotypes at all core loci — against which laboratories calibrate their instruments and validate their allele calls. NIST Standard Reference Material 2391 provides the benchmarks that let a result in Virginia be compared with a result in Oregon. Without that anchor, the numerical output of capillary electrophoresis means something different in each building.

CROSS-REF

The agency sits within a broader standards architecture. SWGDAM — the Scientific Working Group for DNA Analysis Methods — publishes the interpretation guidelines that laboratories work to, covering thresholds, mixture analysis and low-template protocols. ENFSI performs the same coordinating function across European member laboratories, and the Forensic Science Regulator sets statutory quality requirements in England and Wales. But on the question of what the reference allele frequencies are, and which materials calibrate the instruments, NIST is the institutional anchor.

The expansion and its consequences

Expanding a core locus set is not a purely technical decision. Existing profiles in the database were typed at the old thirteen; the new loci had to be added prospectively, and older samples needed re-examination if they were to carry a full twenty-locus profile. Manufacturers of STR multiplex kits — the commercial assays that amplify all core loci in a single reaction — had to validate new products and obtain approval before laboratories could adopt them. The transition took years rather than months.

The expansion also raised the discrimination power of the system substantially. Published estimates suggest the random match probability at twenty loci reaches figures so small they are effectively negligible in most forensic contexts ↗, though analysts are trained to present statistics carefully and not to overstate certainty. The number of loci does not remove the interpretive questions that arise in mixtures, low-template samples or transfer scenarios — it simply means that where a single-source profile can be obtained, the numbers behind it are more robust than ever.