How Investigators Identified the Golden State Killer Through a Family Tree

For decades, the man responsible for a long series of murders, sexual assaults, kidnappings and burglaries in California remained unidentified. The Golden State Killer genetic genealogy investigation changed that, but not through one miraculous database search. It worked because biological evidence had been preserved, forensic scientists could turn that evidence into a genealogy-compatible profile, researchers could build family trees from possible relatives, and detectives could test the resulting lead against DNA obtained from a specific man.

That sequence matters. Genetic genealogy did not convict Joseph James DeAngelo Jr., and a distant relative’s DNA did not establish that he committed the crimes. The genealogy work generated and narrowed a lead. Investigators then obtained discarded DNA associated with DeAngelo and compared it with semen recovered from crime scenes. The documented outcome came later: an arrest, guilty pleas to 26 felony counts, admissions covering additional crimes, and a sentence designed to ensure that he would die in prison.

The investigation began with evidence that survived

The technological breakthrough depended on a decidedly nontechnical decision made years earlier: evidence was retained. Old biological material can be consumed, degraded, contaminated, misplaced or discarded. When it remains available and sufficiently intact, later forensic methods may be able to answer questions that could not be answered when the crime occurred.

The Los Angeles Times reported that a still-sealed sexual-assault kit from the March 13, 1980 killings of Lyman and Charlene Smith in Ventura County supplied semen evidence central to the genealogy effort. The kit connected the later technology to an actual crime and to two people whose lives had been taken. Without preserved biological evidence, there would have been no offender profile to take into a genealogy search.

Preservation alone did not reveal a name. It kept the possibility of identification alive. Investigators still had to extract usable genetic information, interpret it and determine what kind of search that information could support.

Why an ordinary forensic DNA search was not enough

Conventional forensic DNA databases and consumer-oriented genealogy databases answer different questions. The distinction is essential to understanding why the case remained unsolved despite the existence of crime-scene DNA.

Traditional forensic systems such as the FBI’s Combined DNA Index System, commonly called CODIS, use short tandem repeat markers, or STRs. These markers are highly useful for comparing one forensic profile with another. If an unknown crime-scene profile matches a known offender profile or evidence from another case, investigators may obtain a direct lead. But the system cannot name someone whose qualifying profile is not present merely because one of that person’s distant relatives appears somewhere else.

Investigative genetic genealogy instead uses information from large numbers of single-nucleotide polymorphisms, or SNPs, distributed across the genome. Shared stretches of DNA can indicate that two people are biological relatives, even when they are not close relatives. The National Human Genome Research Institute’s explanation of investigative genomics describes the method as a combination of crime-scene DNA analysis, searches for genetic relatives, genealogical research and conventional investigation.

A useful shorthand is that an STR search asks whether the forensic profile matches another profile, while a SNP-based genealogy search can ask whether an unknown person appears to share ancestry with database users. That second question can open a path toward a name, but it also introduces uncertainty. A genetic relative may sit on a family tree containing many branches and hundreds of living people.

What the Golden State Killer genetic genealogy investigation actually did

The genealogy search produced relatives, not an identification. Investigators and genealogical researchers then had to determine how those people might be related to the unknown source of the crime-scene DNA. That meant building family trees backward toward shared ancestors and forward through descendants.

This is where records, rather than DNA alone, became decisive. Genealogical reconstruction can draw on birth, marriage, death and other public records to place people in the correct branches. Conventional investigative work can then test whether people on those branches remain plausible candidates. A family connection creates a pool for examination; it does not establish guilt.

The public story has often been compressed into the claim that investigators uploaded DNA to GEDmatch and found the Golden State Killer. Later Los Angeles Times reporting described a more complicated process. Its reconstruction placed planning and database-search activity by December 2017 and reported activity involving GEDmatch, FamilyTreeDNA and MyHeritage. That account should be understood as the newspaper’s reconstruction of the inquiry, not as a judicial finding establishing every search and procedural detail.

The difference is more than historical trivia. Treating a database as if it returned DeAngelo’s name erases the most laborious part of the process and overstates what the genetic data proved. Relative matching created possible lines of descent. Family-tree research organized those lines. Investigative screening narrowed the people who warranted closer attention. Only then did DeAngelo become a candidate whose DNA could be compared directly with the forensic evidence.

A candidate still had to become an evidentiary match

Once investigators focused on DeAngelo, genealogy had done its principal job. It had supplied a direction for the investigation. Detectives still needed evidence that could connect this particular person—not simply members of his extended family—to the biological material left by the offender.

Released arrest-warrant materials, as reported by the Los Angeles Times, describe a compressed surveillance sequence in April 2018. On April 18, detectives collected DNA from the door of DeAngelo’s car while he was inside a Hobby Lobby. On April 23, investigators recovered a tissue from his trash. The recovered DNA was compared with semen from Golden State Killer crime scenes before his arrest on April 24.

Date	Investigative step	What the step could establish
April 18, 2018	Detectives collected DNA from DeAngelo’s car door during surveillance.	It provided biological material associated with the identified candidate.
April 23, 2018	Investigators recovered a DNA-bearing tissue from DeAngelo’s trash.	It supplied additional material for forensic comparison.
Before April 24, 2018	Investigators compared recovered DNA with semen from Golden State Killer crime scenes.	The direct comparison linked the candidate’s DNA to the crime-scene evidence described in the warrant reporting.
April 24, 2018	Authorities arrested DeAngelo.	The investigation moved from a covertly tested lead to a criminal prosecution.

The car-door and trash collections were fundamentally different from the genealogy search. A relative match suggests where an unknown person might fit within a biological family. A direct comparison tests whether DNA attributed to the candidate is consistent with the crime-scene profile. That final comparison was the corroborating forensic step needed to move beyond family association.

Discarded-DNA collection also allowed detectives to conduct the comparison without first asking DeAngelo to provide a sample. The supplied public record documents what investigators collected and how the material figured in the warrant chronology. It does not provide a complete public laboratory file for every extraction, profile generation, quality-control decision and match calculation made during the broader investigation.

The arrest did not produce a genealogy trial

DeAngelo’s arrest began an enormous prosecution involving crimes committed across multiple California counties. It did not lead to a trial in which jurors heard and evaluated every step of the genealogy operation.

On June 29, 2020, DeAngelo pleaded guilty to 13 counts of first-degree murder and 13 counts of kidnapping to commit robbery. He also admitted additional uncharged crimes. The Orange County District Attorney said prosecutors accepted the agreement after consulting victims and families and considering DeAngelo’s age, the age or death of witnesses, and the burden of pursuing a prolonged capital case.

The prosecution’s sentencing memorandum stated that DeAngelo pleaded guilty to 26 felony counts and admitted violent crimes against 87 victims on 53 occasions in 11 counties. Those are the prosecution’s documented plea and admission figures. They avoid the confusion that can result when accounts combine murders, sexual assaults, kidnappings, burglaries and suspected incidents into one total.

On August 21, 2020, the court sentenced DeAngelo to 11 consecutive life terms without the possibility of parole, an additional consecutive life term for kidnapping, and an eight-year determinate term for weapon enhancements.

The plea secured convictions and punishment without requiring aging victims and witnesses to endure a long capital trial. It also meant there was no complete trial record in which defense experts challenged the genetic-genealogy process, investigators explained each operational decision under cross-examination, and a judge or jury made findings about every disputed part of that process.

What the case established—and what it left unresolved

The investigation established the practical power of combining preserved evidence with genetic genealogy. An unidentified crime-scene profile could be used to locate possible relatives. Researchers could reconstruct family relationships. Detectives could narrow the resulting network and conduct surveillance. A direct forensic comparison could then test whether a candidate was the source of the crime-scene DNA.

The public record does not justify a broader claim that every reported database search was judicially approved or that the case settled all questions about genetic privacy. DeAngelo’s guilty pleas resolved the criminal charges without producing a final courtroom ruling on every genealogy practice described in later reporting.

That distinction is particularly important because database users can expose information about biological relatives who never submitted DNA themselves. A person may choose to participate in genealogy research, yet the resulting match can generate investigative attention toward an extended family. The Golden State Killer case demonstrated the method’s reach before public rules, user expectations and law-enforcement policies had fully caught up with it.

Responsible scrutiny does not require minimizing what the investigation accomplished. It requires describing each step accurately. The relative matches were leads. The family trees organized those leads. Conventional investigation narrowed the candidates. The DNA recovered during surveillance enabled a direct comparison. The guilty pleas and sentence, not the family tree by itself, supplied the case’s legal resolution.

The decisive lesson was the chain, not a single breakthrough

The identification of Joseph James DeAngelo Jr. is often presented as the moment genetic genealogy solved a notorious cold case. The fuller record shows a chain of dependent decisions. Someone preserved the evidence. Scientists obtained different kinds of genetic information from it. Genealogy searches indicated biological relationships. Researchers converted those relationships into family structures. Detectives focused on a candidate and obtained his discarded DNA. Forensic comparison then connected that candidate to the crime-scene evidence described in the warrants.

Remove any major link and the outcome changes. Without retained evidence, there is no profile. Without genealogy, there may be no route to DeAngelo. Without family research and investigative filtering, the database results remain a mass of relatives. Without direct comparison, the family tree remains a lead rather than an identification.

That is what the record can explain with confidence. Genetic genealogy transformed old biological evidence into a path investigators could follow, but disciplined corroboration made the path evidentially meaningful. The case’s enduring lesson is not that a website named a killer. It is that preserved evidence, a new search method, painstaking family research and conventional forensic confirmation worked together to identify one man—and finally brought a decades-long investigation to a documented legal end.