Genetic genealogy has transformed cold case investigations by allowing detectives to identify suspects and victims through their relatives’ DNA, even when the person in question has never submitted a sample to any database. The technique works by comparing crime scene DNA against public genealogy databases, finding distant cousins, and then using traditional family tree research to narrow the search to a single individual. Since its dramatic public debut with the arrest of the Golden State Killer in 2018, forensic investigative genetic genealogy has been applied to hundreds of cases, some of which sat unsolved for decades.
Why Traditional DNA Databases Hit a Wall
For years, forensic labs relied on short tandem repeat (STR) profiles, the kind stored in law enforcement databases like CODIS in the United States. An STR profile works like a fingerprint: if the suspect’s profile is already in the database, the system flags a match. The problem is obvious. If the person who committed the crime was never arrested, never gave a DNA sample, and never entered the system, the search returns nothing. Cold cases pile up precisely because the perpetrator’s profile is absent.
STR analysis also struggles with degraded or low-quality samples, the kind recovered from old crime scenes, buried remains, or evidence stored for years in less-than-ideal conditions. A study comparing next-generation sequencing of single nucleotide polymorphisms (SNPs) against traditional STR typing on degraded skeletal remains found that useful genetic data was far more likely to be recovered using SNP-based methods, especially from samples where STR profiles came back partial or empty.1PubMed. Comparison of next generation sequencing (NGS) – (SNPs) and capillary electrophoresis (CE) – (STRs) in the genetic analysis of human remains SNPs are much shorter fragments of DNA, so they survive degradation better than the longer STR sequences that traditional forensic labs look for.
This shift matters enormously for cold cases. Evidence sitting in a storage room for twenty or thirty years has often deteriorated. In one case involving brutal homicides from 2004, the original DNA extracts contained too little material for standard autosomal analysis. Massively parallel sequencing of mitochondrial DNA allowed the lab to detect contributions as low as three percent in a mixture, connecting the cases to a suspect.2PubMed. From cold case to conviction: How advanced DNA technologies such as mtDNA sequencing connected two brutal homicides Without that newer technology, the evidence would have remained scientifically useless.
From Crime Scene DNA to a Family Tree
Forensic genetic genealogy flips the investigative logic. Instead of asking “is this person in our database?” it asks “who are this person’s relatives?” The process begins by generating a dense SNP profile from the crime scene sample. Dense SNP testing can associate relatives as distant as seventh-degree or beyond, meaning the technique can identify connections through fifth cousins or even more remote relationships.3PubMed Central. Dense single nucleotide polymorphism testing revolutionizes scope and degree of certainty for source attribution in forensic investigations
Once the SNP profile is ready, investigators upload it to a public genealogy database that permits law enforcement use. The database returns a list of genetic matches, each with an estimated degree of relatedness based on shared DNA segments. A genetic genealogist then builds family trees for the closest matches, working backward through birth records, census data, marriage certificates, and obituaries. The goal is to find where two or more family trees intersect, because someone at that intersection could be the source of the crime scene DNA.4PubMed. Investigative genetic genealogy for human remains identification
This genealogical legwork is often the most time-consuming part. A third cousin match might share great-great-grandparents with the unknown person, meaning the genealogist has to trace several generations of descendants to identify living individuals who fit the right age, sex, and geographic profile. When the list is narrowed to one or a handful of candidates, investigators obtain a confirmatory DNA sample through legal means, such as a discarded cup or a cheek swab obtained with a warrant, and run a standard forensic comparison. The genealogy provides the lead; traditional forensic science confirms it.
The Golden State Killer and What Followed
The case that brought forensic genetic genealogy into public view was the arrest of a suspect in California’s Golden State Killer investigation in April 2018. The perpetrator had been linked to at least thirteen murders and over fifty sexual assaults spanning the 1970s and 1980s but had evaded identification for decades. Investigators generated a DNA profile from crime scene evidence and uploaded it to a free online genealogy database used by people researching their family trees. Rather than searching for the killer directly, they searched for his relatives, and then used traditional genealogy to narrow the haystack until a single suspect emerged.5PLoS Biology. Should police have access to genetic genealogy databases? Capturing the Golden State Killer and other criminals using a controversial new forensic technique
The arrest was a watershed moment. Within months, cold case units across the country began adopting the same approach. A profile of solved cases found that many had been open for years or even decades, during which the offenders had continued to victimize others. The same dataset revealed something else striking: cases ultimately solved by genetic genealogy had a meaningful rate of prior wrongful convictions. Roughly three to four percent of the homicides in the dataset had previously led to a wrongful conviction, and wrongful prosecution rates were even higher, reaching over sixteen percent for certain categories of violent crime.6PubMed. Forensic genetic genealogy: A profile of cases solved In at least two instances, people were exonerated as a direct result of a genetic genealogy investigation identifying the actual perpetrator.
Identifying Unidentified Remains
Genetic genealogy is not limited to catching criminals. A large and growing application involves putting names to unidentified human remains. Across the United States, thousands of sets of remains sit in medical examiners’ offices with no identity attached. Some are victims of violence; others died from accidents, natural causes, or drug overdoses. Traditional methods of identification, dental records, fingerprints, and STR database searches, fail when no match exists in any reference database. A pilot study in Australia explored using the same genealogy-based microarray approach that has worked in the United States for reconciling unidentified remains with long-term missing persons.7Forensic Science International. Forensic genetic genealogy using microarrays for the identification of human remains: The need for good quality samples – A pilot study
The challenge with skeletal remains is DNA quality. Bones that have been buried, submerged, or exposed to the elements for years yield DNA that is heavily fragmented. Adapted extraction methods have made progress. Work on highly degraded skeletal remains from the Vietnam War recovered usable autosomal DNA from about seventy percent of tested bone samples, supporting the feasibility of large-scale identifications even from remains that are many decades old.8PubMed. DNA preservation in highly degraded skeletal remains from the Vietnam war-Characterization and DNA extraction evaluation for nuclear SNP panel testing For families who have waited generations for answers, these advances represent something that traditional forensic tools simply could not deliver.
The Privacy Tension
Forensic genetic genealogy depends on public databases populated by ordinary people who uploaded their DNA to learn about their heritage or find biological relatives. Most of those people never imagined their data being used in a criminal investigation, and they certainly didn’t consent on behalf of their entire extended family. When you submit your DNA to a genealogy service, you are making a decision not just for yourself but for every biological relative you have, including those who might prefer to remain invisible to law enforcement.
The legal landscape is genuinely unsettled. Under the third-party doctrine, information voluntarily shared with a third party, such as a direct-to-consumer genetics company, may not carry a reasonable expectation of privacy under the Fourth Amendment.9Houston Law Review. Keeping It in the Family: Direct-to-Consumer Genetic Testing and the Fourth Amendment Police have frequently used consumer genomic platforms without judicial oversight, uploading profiles and searching for matches without obtaining a warrant first.10PubMed Central. Familial Searches, the Fourth Amendment, and Genomic Control
The database companies have responded, if unevenly. In 2019, both FamilyTreeDNA and GEDmatch updated their terms of service and privacy policies to let users opt out of having their data shared with law enforcement.11Nature Biotechnology. Forensic genealogy and the power of defaults The critical detail, though, is the default setting. A policy that makes users opt out rather than opt in means the majority of profiles remain searchable by police simply because most people never change their settings. The framing of that default has outsized consequences for how much of the population is effectively covered by law enforcement searches.
How Other Countries Are Handling It
The technique originated in the United States, but its use is spreading. Sweden became the first country outside North America to use forensic investigative genetic genealogy to solve a crime in 2019. However, the Swedish Privacy Protection Authority blocked further use in 2021, citing a lack of legislative support and concerns about sensitive personal data being transferred to servers in another country.12PubMed Central. Legislation for forensic investigative genetic genealogy in Sweden Sweden’s experience illustrates a tension that plays out globally: the technology works, but existing privacy laws were written before anyone imagined it. Countries that want to use genetic genealogy are forced to either shoehorn it into outdated legal frameworks or write new legislation, and neither option is simple.
Australia, the United Kingdom, and several European countries have explored or piloted the approach. Each jurisdiction faces its own mix of data protection laws, constitutional traditions, and public attitudes. In places with strong data protection regimes like the EU, the transfer of genetic profiles to U.S.-based genealogy databases raises specific legal obstacles around international data flows. The result is a patchwork: some countries are enthusiastically adopting the method, others are cautiously experimenting, and a few have put it on hold pending legislative action.
DNA Phenotyping as a Complementary Tool
When genetic genealogy cannot produce a match, perhaps because the database returns no close enough relatives, investigators sometimes turn to forensic DNA phenotyping. This approach predicts externally visible characteristics from a DNA sample: hair color, eye color, skin pigmentation, facial structure, and even biogeographic ancestry. Validated tools already exist for predicting several appearance traits along with ancestry, and research into age estimation from DNA is advancing.13PubMed. Recent advances in Forensic DNA Phenotyping of appearance, ancestry and age
Phenotyping does not identify a specific individual the way genealogy matching does. Instead, it narrows the suspect pool by describing what the person looks like. In practice, investigators might use a phenotype prediction to guide their genealogical search or to prioritize candidates when the family tree research produces multiple possibilities. The combination of genealogy and phenotyping is more powerful than either alone, but phenotyping also raises its own ethical questions, particularly around predictions of ancestry and the risk of reinforcing racial profiling. Some researchers have argued that all three forms of forensic DNA phenotyping, predicting visible characteristics, ancestry, and age, require proper regulatory frameworks and should be used together rather than in isolation.14PubMed Central. The Use of Forensic DNA Phenotyping in Predicting Appearance and Biogeographic Ancestry Others emphasize that the technology has outpaced the law, and that proportionate regulation and judicial clarity are needed before it becomes routine.15Egyptian Journal of Forensic Sciences. Forensic DNA phenotyping: the need for proportionate regulation and judicial clarity in law enforcement
Whether the Investment Pays Off
Building the laboratory infrastructure for large-scale SNP analysis is expensive. Agencies considering the investment naturally ask whether the results justify the cost. A cost-benefit analysis estimated that a nationwide investment of less than one billion dollars per year over a ten-year period could yield over 4.8 billion dollars annually in tangible and intangible benefits, including solved cases, prevented future crimes, reduced wrongful convictions, and closure for victims’ families. Even if the cost estimates were doubled or tripled, the analysis still projected substantial net benefits.16PubMed Central. A cost-benefit analysis for use of large SNP panels and high throughput typing for forensic investigative genetic genealogy
Those numbers are projections, not guarantees, and they depend on assumptions about how many cases the technology would solve. Still, the finding points to something important: even under pessimistic cost assumptions, the sheer volume of unsolved violent crime in the United States means that a technology capable of solving even a fraction of those cases generates enormous value. The intangible benefits, families receiving answers, wrongfully convicted individuals being freed, and serial offenders being stopped, are harder to quantify but arguably matter even more than the dollar figures.
Security Vulnerabilities in Genealogy Databases
The databases that make forensic genetic genealogy possible are also vulnerable. Research into the security of popular genetic genealogy services has found that these platforms are susceptible to several types of attacks, including genotype extraction, where an adversary retrieves another user’s genetic data, and forged relative attacks, where falsified genetic profiles are uploaded because the platforms do not authenticate the biological origin of submitted data.17University of Washington ResearchWorks. Securing the Future of Biotechnology: A Study of Emerging Bio-Cyber Security Threats to DNA-Information Systems
A forged relative attack is particularly concerning in a forensic context. If someone could upload a fabricated DNA profile designed to appear related to a crime scene sample, they could theoretically misdirect an investigation toward an innocent family. Conversely, an adversary could upload profiles designed to obscure genuine matches. These are not merely theoretical risks; they follow from the fundamental design of open genealogy databases, which were built for hobbyists, not for forensic reliability. As law enforcement relies more heavily on these platforms, the gap between their intended purpose and their forensic use becomes a real security problem.
What Genetic Genealogy Cannot Do
For all its power, the technique has clear limitations. It requires a DNA sample of sufficient quality to generate a dense SNP profile, which is not always possible with old or badly degraded evidence. It requires that at least one relative of the person in question has uploaded DNA to a searchable database, and for some populations, database coverage is thin. People with small families, recent immigrants from underrepresented regions, or members of communities that are less likely to use consumer genetics services may simply have no findable relatives in the system.
The genealogical research phase can also hit dead ends. Adoption, name changes, non-paternity events (where the biological father differs from the assumed father), and poor historical record-keeping all complicate the process. A genetic genealogist might identify a likely branch of a family tree but be unable to resolve which individual within that branch is the right one. And the whole process depends on the accuracy of public genealogical records, which are maintained by amateur researchers of varying skill and thoroughness.
There is also a temporal limitation that gets less attention. Genetic genealogy works best when there is a known set of crimes tied to a single unidentified perpetrator, giving investigators a clear target profile to upload. It is less straightforward when the question is murkier, for instance, when a cold case lacks physical evidence entirely, or when the available DNA belongs to the victim rather than the suspect. The technique opens many doors, but it is not a skeleton key for every unsolved crime.
The Accuracy of Relationship Predictions
When a genealogy database returns a match, it estimates the degree of relationship based on how much DNA two people share. Those estimates are probabilistic, not exact. Two people who share a certain amount of DNA might be second cousins, or they might be first cousins once removed, or they might have a different relationship entirely that happens to produce a similar amount of shared genetic material. Research into the probability distributions of shared DNA across a wide range of relationships has shown that the statistical models used for these predictions can be improved, particularly for non-lineal relationships like cousins and half-siblings, where existing methods sometimes produce inaccurate results.18arXiv. Likelihood Models for Forensic Genealogy
For investigative purposes, some imprecision in the relationship estimate is tolerable. If a match is identified as somewhere between a second and third cousin, the genealogist simply has to explore a slightly wider swath of the family tree. But when cases go to court, the precision of these estimates matters more. Defense attorneys can and do challenge the statistical foundations of genetic genealogy, and the field is still developing the rigorous probabilistic frameworks that courts demand. The science works well enough to generate investigative leads, but the transition from “lead” to “evidence” requires the kind of statistical validation that is still catching up with the technology’s rapid deployment.