How Large Should a Forensic DNA Population Database Be? New 8,237-Person Study Tests the Numbers

How many people are enough for a forensic DNADNA, or Deoxyribonucleic Acid, is the genetic material found in cells, composed of a double helix structure. It serves as the genetic blueprint for all living organisms. Read Full Definition population database? A new study published in the International Journal of Legal Medicine on October 6, 2026, gives one of the clearest empirical answers yet: it depends on what you expect the database to do.

Researchers analyzed 8,237 unrelated individuals from Argentina across 22 autosomal STR lociA locus is the precise position of a gene on a chromosome. Different alleles of the same gene occupy the same locus. Read Full Definition and repeatedly resampled the dataset at different sample sizes. Their results show a striking split. Common forensic summary parameters became stable relatively quickly, while recovery of rare allelesAlleles are different forms of a gene resulting from mutations or variations in the DNA sequence or gene expression. They can be dominant (expressed with one copy) or recessive (expressed only with two copies). Read Full Definition continued to improve as thousands more individuals were added.

That distinction matters because a population database can look statistically stable while still failing to capture a substantial part of the population’s low-frequency allelic variation.

Important terminology: this article concerns a population allele-frequency database used for forensic statistical interpretation. It is not an offender or criminal DNA database.

Study at a glance

The study, led by Antonella Belén Penacino and colleagues, examined the effect of sample size on allele diversity and commonly used forensic parameters. The full Argentine dataset contained 8,237 individuals genotyped at 22 autosomal STR loci.

Instead of simply publishing allele frequencies from the full cohort, the researchers asked a more practical question: what would the same population database look like if only 500, 1,000, 3,000, 5,000 or several thousand people had been sampled?

To investigate this, they generated 1,000 random resampling replicates across a range of sample sizes extending from 500 individuals to the full cohort. This allowed them to observe how quickly allele recovery and forensic parameters approached the values obtained from the much larger dataset.

The full dataset contained many rare alleles

Across the 22 STR loci, the researchers observed 344 distinct alleles. Of these, 169 alleles, or 49.1%, occurred at frequencies below 1%.

That is a key result. Almost half of the observed alleles belonged to the rare end of the frequencyFrequency is a fundamental concept in physics and wave theory. It refers to the number of times a specific point on a wave, such as a crest or trough, passes a fixed reference point in Read Full Definition distribution.

Rare alleles are difficult to capture in small population samples for a simple reason: an allele that genuinely exists in the population may not appear at all in a modest reference sampleReference sample - material from a verifiable/documented source which, when compared with evidence of an unknown source, shows an association or linkage between an offender, crime scene, and/or victim. Read Full Definition. Its absence from the database does not mean it is absent from the population.

This is also why forensic laboratories use approaches such as minimum allele-frequency rules when an observed allele has not been represented adequately in the reference dataset.

Overall allele recovery improved rapidly, but not immediately

The study found that increasing sample size progressively recovered more of the allelic diversity present in the complete dataset.

A sample of 500 individuals therefore recovered roughly three-quarters of the total allelic diversity observed in the 8,237-person dataset. At 3,000 individuals, overall recovery exceeded 90%, and by 7,000 it approached saturation.

But the rare alleles followed a slower curve.

Rare alleles were much harder to capture

At a sample size of 3,000, the database recovered only 81.0% of the rare alleles seen in the complete cohort. Even at 5,000 individuals, rare-allele recovery reached only 90.4%.

This is the most important practical message from the study.

If the objective of a population database is simply to obtain stable combined forensic parameters, a moderate sample may appear sufficient. If the objective is to characterize the full allele spectrum, especially low-frequency alleles, the required sample size can be far larger.

In other words, the question should not be:

What is the minimum acceptable sample size for a forensic population database?

A better question is:

What parameter are we trying to estimate, and how precisely do we need to estimate it?

Combined forensic parameters were surprisingly stable

The researchers also evaluated combined forensic parameters, including combined match probability and combined power of exclusion.

These measures showed comparatively little variation across the different sample-size scenarios. Even smaller resampled datasets produced combined values close to those calculated from the full cohort.

This happens because multilocus forensic statistics combine information across many STR loci. Small changes in individual allele-frequency estimates do not necessarily produce large changes in an already highly discriminating multilocus system.

For readers who want a broader explanation of how population frequencies enter forensic DNA statistics, see our guide to the likelihood ratio in forensic statistics and our overview of DNA evidence interpretation in criminal cases.

Stable statistics do not mean the database is complete

This is where the study becomes particularly useful.

A database can produce stable combined match or exclusion statistics and still be missing a meaningful proportion of rare alleles.

Those are two different measures of adequacy.

Statistical stability asks whether the calculated forensic parameter changes materially when more individuals are added.

Allelic representation asks whether the database has captured the range of variants actually present in the population.

The Argentine dataInformation in analog or digital form that can be transmitted or processed. Read Full Definition show that the first can be achieved much earlier than the second.

Why unseen alleles matter

Suppose an allele appears in a case sample but has not been observed in the reference database. It would be inappropriate to assign that allele a population frequency of zero.

The fact that the allele was not sampled does not establish that it does not exist in the population. Instead, laboratories apply conservative statistical procedures, including minimum-frequency approaches, according to their validated interpretation framework and applicable guidance.

Larger and more representative population datasets reduce the chance that genuine low-frequency alleles remain unobserved and can improve the empirical foundation for allele-frequency estimates.

The underlying concept is straightforward: rare events require more observations before they can be characterized reliably.

This does not mean every forensic database needs 8,000 people

The study should not be interpreted as establishing 8,237, 7,000 or even 5,000 as a universal minimum sample size.

The results came from one large Argentine cohort and 22 autosomal STR loci. The sample size required in another population will depend on factors such as:

• the loci being examined;
• the number and frequency distribution of alleles;
• population history and substructure;
• sampling design and geographic representation;
• the forensic parameter being estimated;
• whether the objective is routine frequency estimation or extensive rare-allele characterization.

The authors also found that loci with a larger proportion of rare alleles took longer to approach saturation. This makes a single universal sample-size threshold difficult to justify scientifically.

What this means for forensic population studies

The paper supports a more defensible way of designing and evaluating forensic population datasets.

Researchers should define the purpose of the database before deciding whether the sample is sufficiently large. A study intended to demonstrate stable combined forensic parameters may reach that goal with a smaller cohort than a study intended to document low-frequency allelic diversity comprehensively.

For population-genetics publications, this also means that reporting only high discrimination or exclusion values can hide an important part of the picture. Authors should consider whether allele discovery is approaching saturation and how many low-frequency or singleton alleles remain sensitive to sample size.

Representativeness remains just as important as raw sample count. A very large but geographically or demographically biased sample does not automatically provide a better population reference than a smaller, carefully designed dataset.

Why the study matters

Forensic genetics has long relied on population reference data, but sample-size decisions are often justified using convention, precedent or practical feasibility. This study provides unusually large empirical data showing exactly how different aspects of a forensic STR database respond as the number of sampled individuals increases.

The answer is not that small databases are useless, nor that every population study must contain thousands of individuals.

The answer is more precise"Precise" refers to the degree of closeness or consistency between multiple measurements or values taken under the same conditions. It indicates how well these measurements agree with each other, regardless of whether they are accurate Read Full Definition:

A population database can become adequate for one forensic purpose before it becomes adequate for another.

Combined statistics may stabilize relatively early. Rare-allele representation does not.

That is a distinction worth making whenever researchers, laboratories or reviewers ask the deceptively simple question: Is this population sample large enough?

Reference

Penacino AB, Carvajal-Pérez CE, Rangel-Villalobos H, Elsztein LD, Puentes PA, Zapata FA, Becerra-Loaiza DS, Moreno-Ortiz JM, Penacino GA, Aguilar-Velázquez JA. Sample size effects on allele diversity, rare-allele recovery, and forensic parameters of 22 autosomal STRs in a cohort of 8,237 Argentines. International Journal of Legal Medicine. Published October 6, 2026. DOI: 10.1007/s00414-026-04032-4.