Short tandem repeat analysis is the forensic DNA method that identifies a person by counting how many times a short genetic sequence repeats itself at 20 standardized locations in their DNA. Those 20 repeat counts, recorded as pairs, produce a numerical profile so specific that the odds of an unrelated person sharing it are effectively zero. It is the science behind criminal DNA matches, paternity testing, disaster victim identification, and the FBI’s national DNA database.
What an STR Actually Is
Human DNA contains regions where a short sequence of two to six chemical building blocks repeats back to back. These stretches are called short tandem repeats. Everyone inherits two copies of each region, one from each parent. The repeating sequence itself is common across the population, but the number of repeats varies from person to person. One individual might carry 10 repeats at a given location while another carries 15 at that same spot.
These regions sit in non-coding stretches of DNA, so they don’t carry instructions for building proteins. That makes them stable markers: they stay the same throughout your life and read the same whether the sample comes from blood, saliva, or skin.1National Center for Biotechnology Information (NCBI). The Evaluation of Forensic DNA Evidence Recording the repeat count at enough locations and stacking those numbers together produces a profile unique to one person. The only people who share an STR profile are identical twins.
The 20 CODIS Core Loci
For a DNA profile to enter the national database, the laboratory has to test a standardized set of genetic locations. Since January 1, 2017, that set includes 20 locations, known as the CODIS core loci. Before that date, only 13 were required.2Federal Bureau of Investigation. CODIS and NDIS Fact Sheet Expanding from 13 to 20 markers cut the already tiny probability of a coincidental match by orders of magnitude and made it easier to share data with international databases that test overlapping markers.
Each locus carries a standardized name (like D13S317 or FGA) tied to its chromosome location. A complete profile records two allele values at each locus, producing a string of 40 numbers. When analysts compare two profiles, they compare those 40 numbers. A single discrepancy at one locus rules out a match.
How the Lab Turns a Sample Into a Profile
Forensic DNA work starts with the sample itself. Common biological sources include blood, cheek swabs, skin cells, hair with the root attached, and saliva.1National Center for Biotechnology Information (NCBI). The Evaluation of Forensic DNA Evidence Wet biological evidence has to be air-dried before packaging, because sealing damp material in plastic promotes mold that destroys the DNA entirely. Dried items go into sealed paper bags or envelopes, packaged separately to prevent cross-contamination.3National Institute of Justice. Proper Evidence Collection and Packaging
Every sample also has to travel under a documented chain of custody, a written record of everyone who handled the evidence from collection through analysis and storage. Each transfer is logged by name, date, and time. A gap in that chain can get the evidence excluded at trial or reduce the weight the jury gives it.4National Institute of Justice. Chain of Custody Cases fall apart at this stage far more often than at the science stage.
Once the sample reaches the lab, technicians isolate the DNA from proteins, cell membranes, and chemical inhibitors that could interfere with testing. The purified DNA then goes through polymerase chain reaction, or PCR, which makes millions of copies of just the 20 target STR regions. A thermal cycler rapidly heats and cools the sample through dozens of cycles, each one doubling the amount of target DNA.
After amplification, the copied fragments are sorted by size through capillary electrophoresis. The DNA is injected into a hair-thin glass tube and pulled through a gel-like substance by an electric current. Because DNA carries a negative charge, the fragments migrate toward the positive end of the tube. Smaller fragments travel faster, so they separate by length as they go.
As each fragment passes a detection window, a laser strikes fluorescent dye tags attached during PCR. Software captures the light pulses and plots them as peaks on a chart called an electropherogram. Each peak corresponds to an allele at one of the 20 loci, and its position on the chart indicates how many repeats that allele contains. The software converts those peaks into the numerical profile that gets compared against other profiles or entered into a database.
How Strong a DNA Match Actually Is
A DNA profile by itself is just a string of numbers. Its evidentiary power comes from the statistical calculation attached to it. When two profiles match at every tested locus, the laboratory calculates the probability that a randomly selected, unrelated person from the population would share that same profile. This figure is called the random match probability.
The calculation multiplies the individual frequencies of each allele pair across all tested loci, with corrections for population structure. Because allele frequencies at each locus are largely independent, combining 20 loci produces astronomically small probabilities. A full 20-locus match routinely yields a random match probability smaller than one in a trillion, which effectively rules out a coincidental match among unrelated people. Under the older 13-locus system the numbers were already extremely small; the expansion to 20 made them orders of magnitude smaller.
Partial profiles produce weaker statistics because fewer loci contribute to the calculation. A profile with results at only eight of 20 loci might yield a random match probability of one in several million rather than one in trillions. That is still strong evidence, but it leaves more room for a defense challenge.
When Samples Get Messy: Partial Profiles, Mixtures, and Touch DNA
Not every sample produces a clean, complete profile. Degraded DNA, exposure to heat or sunlight, or a very small starting quantity can cause some loci to drop out, producing a partial profile. A partial profile is still useful for comparison, but its statistical power drops because fewer loci means fewer data points separating one person from another.5National Institute of Standards and Technology. DNA Mixture Interpretation: A NIST Scientific Foundation Review
Mixtures are the harder problem. When a sample contains genetic material from two or more people, the lab sees overlapping peaks on the electropherogram and has to figure out which alleles belong to which contributor. The difficulty scales with the number of contributors, the relative amounts of DNA each person left behind, and whether alleles from different contributors happen to overlap. Forensic scientists use statistical models and expert judgment to work through these mixtures, but honest practitioners will tell you that mixtures involving three or more contributors at similar quantities push the limits of reliable interpretation.5National Institute of Standards and Technology. DNA Mixture Interpretation: A NIST Scientific Foundation Review
Modern laboratories increasingly rely on probabilistic genotyping software rather than older binary methods. Instead of declaring each locus “included” or “excluded,” probabilistic models weigh the likelihood of observing the data under different scenarios, accounting for the possibility of dropped alleles and other artifacts. The result is expressed as a likelihood ratio rather than a simple yes or no. Different software packages and different analysts can assign different likelihood ratios to the same mixture, which is why defense experts frequently scrutinize the assumptions behind the calculation.
Touch DNA
Improved sensitivity has made it possible to develop profiles from trace amounts of skin cells left by casual contact, commonly called touch DNA. A profile pulled from a doorknob, a steering wheel, or a weapon handle can generate leads that older technology would have missed.6National Institute of Justice. Persistence of Touch DNA for Analysis The tradeoff is that touch DNA yields low quantities of genetic material with high variability in quality, and there is a real concern about secondary transfer. Your DNA can end up on an object you never touched if it was carried there by an intermediary. Shaking someone’s hand and then having that person handle a piece of evidence could deposit your DNA on it. Both sides need to account for that possibility when evaluating touch DNA results.
Comparing Profiles Through CODIS
The Combined DNA Index System, known as CODIS, is the FBI program that manages forensic DNA databases at the local, state, and national levels. The national tier, the National DNA Index System or NDIS, contained over 19.2 million offender profiles, 6.1 million arrestee profiles, and 1.4 million forensic profiles as of November 2025, and had assisted in more than 758,000 investigations.7Federal Bureau of Investigation. CODIS-NDIS Statistics
Federal law authorizes the FBI Director to establish the national index, which includes DNA records from convicted offenders, arrestees, crime scene samples, and unidentified human remains.8Office of the Law Revision Counsel. 34 USC 12592 – Index to Facilitate Law Enforcement Exchange of DNA Identification Information A separate statute directs the FBI to publish quality assurance and proficiency testing standards that every participating laboratory has to meet.9Office of the Law Revision Counsel. 34 USC 12591 – Quality Assurance and Proficiency Testing Standards Laboratories that fail proficiency tests risk losing their ability to upload profiles.
In practice, a laboratory uploads a crime scene profile to CODIS and runs it against the offender and arrestee indexes. If the software finds a matching profile, it flags the result as a “hit” and identifies the individual whose record produced the match. That hit then goes back to the investigating agency for confirmation through an independent analysis before it can be used in court. CODIS has produced over 781,000 hits since its inception.7Federal Bureau of Investigation. CODIS-NDIS Statistics
Some jurisdictions also conduct familial searching, a deliberate secondary search for profiles that partially match a crime scene sample in a way that suggests a close biological relative of the unknown contributor. The FBI does not perform familial searching at the national level and does not regulate whether states do so. Each state decides independently, and some have passed laws specifically prohibiting the practice.2Federal Bureau of Investigation. CODIS and NDIS Fact Sheet
Where STR Evidence Gets Challenged in Court
Federal courts evaluate scientific evidence under the framework from Daubert v. Merrell Dow Pharmaceuticals (1993), which asks whether the technique has been tested, peer-reviewed, has a known error rate, is governed by standards, and is broadly accepted in the relevant scientific community.10Legal Information Institute. Daubert Standard STR analysis satisfies all of those criteria convincingly, and no federal court has rejected the core methodology. A significant number of states still follow the older Frye standard, which asks only about general acceptance, and STR analysis clears that bar as well.
The productive defense challenges are almost never about whether STR profiling works. They target three things: breaks in the chain of custody, contamination during collection or laboratory processing, and the assumptions behind mixture interpretation. Laboratories respond through proficiency testing, blind quality-control samples, and detailed documentation of every analytical step.
If Your DNA Is Already in the Database
The legal authority to collect DNA depends on the situation. Crime scene evidence is governed by ordinary Fourth Amendment rules. Every state requires DNA collection from convicted offenders. Collection at arrest, before any conviction, is the more contested category. The Supreme Court addressed it in Maryland v. King (2013), holding 5-4 that taking a cheek swab from someone arrested for a serious offense is a reasonable search under the Fourth Amendment, comparable to fingerprinting and photographing during booking.11Legal Information Institute. Maryland v. King The practical result is that most states now collect DNA at the point of arrest for qualifying offenses.
If your DNA profile is entered into the national database after an arrest, an acquittal or dismissed charge does not automatically remove it. You have to request expungement affirmatively. For profiles entered under federal authority, that means a written request to the FBI’s Federal DNA Database Unit in Quantico, Virginia, with a certified copy of the final court order showing the charge was dismissed or resulted in an acquittal. The order has to be signed by a judge and contain enough identifying information (at minimum your full name plus a Social Security number or date of birth) to confirm your identity and the specific charge.12Federal Bureau of Investigation. DNA Fingerprint Act of 2005 Expungement Policy
Expungement means complete removal of the profile from NDIS and destruction of the physical DNA sample. Without the certified court order, the FBI will not process the request. State-level profiles follow separate procedures set by each state’s own laws. Many people who are acquitted or have charges dropped never learn about this process, and their profiles sit in the database indefinitely. If your federal case was resolved in your favor, the expungement request is worth making.