Every other way of catching a killer with DNA needs the killer's own profile. A sample from the scene is compared against a database of offenders, and if he has never been arrested, nothing matches and the case stays open for thirty years.
Seven cases in this archive turned instead on somebody else's DNA. A cousin's, an uncle's, a son's. In five of them it produced a name. In two it has not yet, and those two show exactly where the method stops.
Lonnie Franklin was identified in 2010 through California's first familial DNA search. The state ran the crime scene profile against its offender database, got no match to him, and got a partial match to a profile taken from his son after a weapons conviction. A father and a son share half their DNA, and that was enough to point investigators at the family. An undercover officer working as a restaurant server collected a discarded pizza crust and a napkin, and the DNA on those matched the scenes. He was arrested on July 7, 2010, convicted on May 5, 2016 of ten murders and one attempted murder, and sentenced to death.
That is the older technique and its limit is obvious. It only works if a relative has been convicted of something.
In 2018 investigators uploaded the crime scene profile from a long series of Californian rapes and murders to GEDmatch, an open genealogy database built by hobbyists tracing their ancestry. They found distant cousins, built family trees down from them, narrowed the descendants to one man, and confirmed it from a discarded tissue and a car door handle. Joseph James DeAngelo was arrested at his home in Citrus Heights on April 24, 2018, aged seventy-two. On June 29, 2020, in a converted university ballroom, he pleaded guilty to thirteen murders, and was sentenced to life without parole that August.
The difference between those two cases is the whole subject. Franklin needed a relative with a criminal record. DeAngelo needed a distant cousin with an interest in family history. The second pool is very much larger, and nobody in it consented to anything of the kind.
Terry Peder Rasmussen was convicted once, of the 2002 murder of Eunsoon Jun in California, and died in prison in 2010. In 2017, seven years later, DNA linked him to the four bodies found in barrels near Bear Brook State Park in New Hampshire. The method named the dead as well as the man responsible, which is the part of Bear Brook that investigators cite: it took nearly forty years, and it was finished long after he was beyond any court.
Kenneth Smith was never charged with anything. Three murders in Ontario are attributed to him. The sequence is worth setting out, because it shows how slow this is even when it works. In 2016 Canada's national DNA data bank matched two of the cases to each other. In 2017 the third joined them. In 2022 investigators went to a forensic genetic genealogy company in Texas, and three years after that they had identified close relatives of the man whose DNA it was. On December 3, 2025 he was conclusively identified. He had died in Windsor in 2019, aged seventy-two.
Both men join a problem this archive has written about from the other side. Eight subjects here died in custody before any court ruled, leaving counts that can never be tested. Rasmussen and Smith are the same defect arriving by a different route: identified with more confidence than most convictions carry, and no proceeding possible against either of them. Smith has three killings attributed and no charge, ever.
Randy Steven Kraft was convicted in 1989 of sixteen murders and up to sixty-seven are attributed to him. Cold cases are still being attached: in November 2023 the remains of a teenager believed killed by him nearly fifty years earlier were identified through genetic genealogy, and in May 2025 Oregon State Police named him as the suspect in a 1980 killing beside Interstate 5.
The Santa Rosa Hitchhiker Murders run the other way. At least seven deaths in Sonoma County in the early 1970s, unsolved, attached over the years to Ted Bundy, to the Zodiac and to a series of local suspects, none of it established. In 2021 investigators announced that DNA had connected one of those deaths, Jeannette Kamahele's, to a separate offender.
That is the useful thing about the technique and the reason it deserves its own piece rather than a paragraph inside the DNA reopenings. It does not only add names to a killer. It takes deaths out of a series that never belonged there, and a series is a claim about a single person.
The I-70 Killer has six deaths attributed across Indiana, Missouri and Kansas, with two Texas killings in 1993 linked by ballistics. Thousands of tips, a composite sketch circulated for thirty years, a nationally publicised push in 2019, and genetic genealogy work since. The case is open and the FBI still lists it.
Santa Rosa is the same, and both cases fail for the same reason. Genealogy needs a relative who has uploaded a profile to a consumer site. Where nobody in that family has, there is no tree to build down. The crime scene DNA sits in a file being perfectly good evidence against a man nobody can name.
The technique reversed something basic. For a century, being unknown to the police was a serial killer's best protection, and it worked for DeAngelo for over forty years. It no longer does, because the thing that identifies him is not his record. It is his family.
But two of the seven were named too late for any court, two are still unnamed because the right cousin never bought a test, and one series has had a death removed from it. Five names in fifty years of casework is not a machine. It is a method that depends entirely on who else in a man's family happened to be curious about where they came from.