DNA evidence is highly reliable in court when the sample is clean, the lab work is sound, and the statistics are honestly presented, but its reliability in any given case is not automatic. A full DNA profile developed under proper conditions can reduce the odds of a coincidental match to less than one in a billion. Contamination, secondary transfer, degradation, mixture interpretation, and misunderstood statistics have all contributed to wrongful convictions and disputed results, and the technology keeps outpacing the legal standards written to govern it.
How Discriminating a DNA Match Can Be
Nearly every person carries a unique DNA sequence. Identical twins are the significant exception; their standard forensic profiles are indistinguishable with current testing. Analysts extract DNA from biological material at a scene, such as blood, saliva, skin cells, or hair roots, and examine short tandem repeats (STRs), regions where a pattern repeats a variable number of times from person to person. Measuring the repeat count at multiple locations yields a numerical profile that functions as a genetic fingerprint.
These profiles run through the FBI’s Combined DNA Index System, or CODIS, which links local, state, and national databases.1FBI. CODIS and NDIS Fact Sheet As of November 2025, the national index held over 19.2 million offender profiles and roughly 1.4 million forensic profiles from unsolved cases.2FBI. CODIS-NDIS Statistics A CODIS hit is an investigative lead, not evidence; it still needs independent corroboration before it can support a conviction.
Where DNA Evidence Goes Wrong
The underlying science is well established. The problem is that every step between a crime scene and a lab report creates room for error. None of the issues below means DNA evidence is unreliable as a category. Each means a specific result deserves questioning.
Contamination
Foreign DNA can enter a sample during collection or lab processing. An investigator who sneezes near evidence, a technician who handles two cases without changing gloves, or a shared workstation that was not properly cleaned can all introduce stray genetic material. Even trace amounts can distort a profile or render it uninterpretable. Labs use protective equipment, single-use tools, and separated workspaces to reduce the risk, but no protocol eliminates it.
Secondary Transfer
Secondary transfer is where DNA evidence gets genuinely tricky and where juries are most easily misled. Your DNA can reach a surface you never touched, carried by an intermediary. Shake someone’s hand, and if that person then picks up a coffee mug, your DNA may end up on the mug. Research has shown DNA from one person’s keyboard transferring to a second user’s hands up to eight days later, and gloves worn during evidence processing can move DNA between exhibits.3NCBI. Indirect DNA Transfer and Forensic Implications: A Literature Review In simulated workplace scenarios, the intermediary’s DNA has appeared as the major contributor in the transferred sample rather than the original source. Even household pets can act as carriers. The practical consequence: finding someone’s DNA at a scene does not prove that person was ever there.
Degradation
DNA breaks down with heat, moisture, sunlight, and bacteria. A bloodstain baked on a sidewalk for days will yield a weaker, more fragmented profile than one collected within hours and stored properly. Degraded samples often produce partial profiles, with results at some STR locations but not others. A partial profile still has value, but it matches a larger portion of the population than a full one, and its statistical weight drops accordingly.
Human Error
Mislabeled tubes, improperly calibrated instruments, and procedural shortcuts have derailed real cases. A sample swapped between two cases during processing looks exactly like a legitimate match until someone catches the paperwork discrepancy. Standardized protocols and proficiency testing reduce these errors, but the defense is entitled to examine whether the safeguards actually held in a particular case.
Mixtures
Crime-scene samples often contain DNA from more than one person. A door handle, a weapon that passed between hands, or a sexual assault kit can all yield mixed profiles. Separating individual contributors is one of the hardest problems in forensic genetics. When one contributor dominates, analysts can sometimes pull out a clean profile. When contributions are roughly equal, or when a minor contributor is present in tiny amounts, interpretation involves judgment calls that different analysts can resolve differently.
What the Statistics Actually Mean
A DNA “match” is never presented as certainty. Analysts express results as statistical probabilities, and understanding those numbers is essential to weighing the evidence honestly.
Random Match Probability
The most common statistic is the random match probability. It answers one question: if you pulled a person at random from the population, what are the odds their DNA profile would coincidentally match the crime-scene profile? An expert might testify that this probability is one in 10 billion, meaning the profile is extraordinarily rare.4NCBI Bookshelf. The Evaluation of Forensic DNA Evidence – DNA Evidence in the Legal System The figures come from population databases tracking how common each STR pattern is across demographic groups.
Likelihood Ratios
Labs increasingly report a likelihood ratio alongside or instead of the random match probability. A likelihood ratio compares two scenarios: how probable is the DNA evidence if the suspect is the source, versus how probable if someone else is? For single-source cases, the likelihood ratio is simply the inverse of the random match probability. If the random match probability is one in a billion, the likelihood ratio is a billion to one in favor of the suspect being the source.5NCBI Bookshelf. The Evaluation of Forensic DNA Evidence – Statistical Issues The likelihood ratio framework matters most with complex mixtures, where probabilistic genotyping software generates a ratio rather than a simple frequency.
The Prosecutor’s Fallacy
This is the single most dangerous statistical mistake in DNA cases. The prosecutor’s fallacy confuses two very different questions: “What is the probability a random innocent person would match this profile?” and “Given the match, what is the probability this defendant is innocent?” They sound similar. They are not the same.
The Supreme Court addressed this in McDaniel v. Brown (2010). If a juror hears a random match probability of one in 10,000 and treats that as a one-in-10,000 chance that someone other than the defendant left the DNA, the juror has fallen for the fallacy. The actual probability that the defendant is the source depends on all the evidence in the case, not the DNA statistics alone. In a city of a million people, a one-in-10,000 random match probability means roughly 100 people would match. The DNA alone doesn’t tell you which one is the right one.
What Judges Require Before a Jury Hears It
Before DNA evidence reaches a jury, a judge decides whether the methods behind it are reliable enough to admit. That gatekeeping role has grown more demanding, not less.
The Daubert Standard
Federal courts and most states follow the framework the Supreme Court laid out in Daubert v. Merrell Dow Pharmaceuticals, Inc. (1993). Judges weigh whether the technique has been tested, whether it has been subjected to peer review and publication, its known or potential error rate, whether standards control its operation, and whether it is generally accepted in the relevant scientific community.6Justia. Daubert v. Merrell Dow Pharmaceuticals, Inc. No single factor is decisive.
The Frye Standard
Several states, including some of the most populous, still use the older test from Frye v. United States (1923). Under Frye, the only question is whether the technique has gained “general acceptance in the particular field in which it belongs.” The judge doesn’t independently evaluate error rates or methodology. DNA profiling itself clears Frye easily because the science has been accepted for decades, but newer interpretive tools like probabilistic genotyping software face closer scrutiny in Frye jurisdictions, where a novel method must show broad consensus before a court will allow it.
The 2023 Amendment to Federal Rule 702
Effective December 1, 2023, the federal rules raised the bar for all expert testimony. Rule 702 now requires the party offering an expert to demonstrate that “it is more likely than not” that the testimony rests on sufficient facts, uses reliable methods, and reflects a reliable application of those methods to the case at hand.7United States Courts. Federal Rules of Evidence The advisory committee said the change responded to courts that had been admitting expert testimony too liberally, particularly in forensic fields where witnesses stated conclusions beyond what the science could support. For DNA, this means federal judges are expected to examine not only whether DNA analysis is generally reliable, but whether the specific analyst applied reliable methods reliably in the specific case.
How the Defense Can Test the Evidence
DNA evidence carries enormous weight with juries, which is why effective challenges matter. Defense attorneys target the evidence at every stage, from collection to interpretation.
Cross-Examining the Analyst
Defense attorneys probe prosecution experts for bias, inconsistencies with prior testimony, deviations from lab protocols, and gaps in qualifications.8National Institute of Justice. Law 101 Legal Guide for the Forensic Expert – Cross-Examination Effective cross-examination often focuses on the assumptions behind statistical calculations, the handling of mixture profiles, and departures from standard operating procedures. The defense doesn’t need to prove the science is junk. It needs to show that this particular result, in this particular case, may not be as airtight as the prosecution suggests.
Proficiency and Error Rates
Forensic labs undergo periodic proficiency testing, and defense attorneys can ask whether the lab or the individual analyst has ever failed one. A failed proficiency test doesn’t automatically invalidate every result the analyst has produced, but it gives the defense a concrete basis for questioning reliability.9National Institute of Justice. Error Rates The defense can also challenge whether a lab’s stated error rate reflects its actual performance, since proficiency tests tend to be easier and cleaner than real casework.
Discovery and Lab Records
Federal rules require the prosecution to turn over the lab’s case file on request, which normally includes the chain-of-custody log, the analyst’s bench notes, photographs, worksheets, data charts, and internal emails between analysts or between analysts and prosecutors. The government must also provide a written summary of any expert testimony, including the expert’s opinions and reasoning. Information giving the defense a basis for challenging test results may qualify as material that must be disclosed under the Constitution.
Access to the proprietary source code behind DNA analysis software has been more contested. Some courts have allowed defense teams to examine code under a protective order. At least one appeals court has held that source code access is unnecessary because validation studies and expert testimony are enough to evaluate reliability. As probabilistic genotyping plays a growing role, these disputes will continue.
Independent Experts
Defense teams can retain their own forensic experts to re-analyze evidence, review the prosecution’s lab work, and offer alternative interpretations. An independent expert might identify problems in mixture interpretation, flag deviations from procedure, or explain the limits of the evidence in terms a jury can follow. Courts have recognized that defense counsel may need expert assistance simply to understand what the prosecution’s reports mean, let alone challenge them.
New Technologies That Change the Reliability Picture
Forensic DNA science keeps moving, and the newest tools raise legal questions courts are only beginning to answer.
Rapid DNA
Rapid DNA instruments can generate a profile from a mouth swab in about 90 minutes, without a traditional lab. The FBI has approved their use at police booking stations for processing arrestee samples and searching them against CODIS, provided the state has a qualifying arrestee DNA collection law and the necessary electronic fingerprint integration.10FBI. Guide to All Things Rapid DNA As of mid-2025, the FBI updated its quality standards to allow crime-scene evidence samples to be processed on Rapid DNA instruments within accredited labs under strict conditions, though booking stations remain limited to arrestee reference samples. Seven states currently operate non-CODIS statewide programs using Rapid DNA to test evidence against state-level databases.
Probabilistic Genotyping Software
Programs like STRmix and TrueAllele use algorithms to interpret DNA mixtures that would be impossible to resolve by hand. Instead of an analyst reading a messy chart of overlapping peaks, the software runs thousands of simulations and produces a likelihood ratio for each possible contributor. Both programs have been validated by multiple crime labs and used in casework nationwide. They are not interchangeable. In one notable case, STRmix excluded a suspect while TrueAllele returned an inconclusive result from the same profile. That kind of divergence undercuts the image of DNA analysis as purely objective. Federal rulemakers are paying attention: a proposed Federal Rule of Evidence 707, under discussion as of early 2026, would address the admissibility of machine-generated inferential evidence, with probabilistic genotyping named as a key example.
Investigative Genetic Genealogy
The technique that identified the Golden State Killer works by uploading a crime-scene profile to a consumer genealogy database such as GEDmatch, finding partial matches among distant relatives, and building a family tree backward to identify a suspect. It has solved cold cases dormant for decades. It also raises privacy questions, because the relatives whose DNA leads investigators to a suspect never consented to being part of a criminal investigation. Courts that have addressed the issue have generally found that people lack a reasonable expectation of privacy in genomic data they voluntarily shared with genealogy services, though the law is still developing and no binding federal appellate ruling has squarely resolved the Fourth Amendment question.
DNA and Wrongful Convictions
DNA’s track record cuts both ways. It is simultaneously the most powerful tool for identifying the guilty and the most effective tool for freeing the innocent. As of mid-2016, DNA analysis had led to the exoneration of 342 wrongfully convicted people in the United States.11National Institute of Justice. Wrongful Convictions and DNA Exonerations – Understanding the Role of Forensic Science The count has continued to grow, with the Innocence Project alone representing over 200 DNA exonerees through early 2026.
Most of those wrongful convictions were not caused by DNA errors. The primary drivers were eyewitness misidentification and older forensic disciplines like microscopic hair comparison and ABO blood typing, which lacked the discriminating power DNA analysis provides. Where forensic science contributed to a wrongful conviction, DNA itself accounted for a small fraction; the overwhelming majority involved serology or hair analysis performed before modern DNA methods existed.11National Institute of Justice. Wrongful Convictions and DNA Exonerations – Understanding the Role of Forensic Science The takeaway is not that DNA is unreliable; it is that other kinds of evidence long treated as strong were weaker than anyone realized.
Post-Conviction DNA Testing
A person convicted of a federal crime can request DNA testing under 18 U.S.C. ยง 3600. The statute sets a high bar. The applicant must assert actual innocence under penalty of perjury, identify specific evidence that was either never tested or could benefit from substantially improved testing, and show the evidence has been preserved under a reliable chain of custody. Identity must have been at issue at the original trial, and the proposed testing must be capable of producing new evidence that would raise a reasonable probability the applicant would not have been convicted.12Office of the Law Revision Counsel. 18 US Code 3600 – DNA Testing
All 50 states have their own post-conviction DNA testing laws, though the specifics vary. Most share common elements: identity was central at trial, the evidence has been properly maintained, and results could have changed the outcome. Some states permit retesting of previously examined evidence only when significantly more accurate technology has become available. These statutes exist because the science keeps advancing. A sample that yielded nothing useful with 1990s technology may produce a full profile today, and a profile that was too degraded for a single-source interpretation then may be resolvable with probabilistic genotyping software now.