BY MEGAN MOLTENI
In April, a citizen scientist named Barbara Rae-Venter used a little-known genealogy website called GEDMatch to help investigators find a man they’d been looking for for nearly 40 years: The Golden State Killer. In the months since, law enforcement agencies across the country have flocked to the technique, arresting a flurry of more than 20 people tied to some of the most notorious cold cases of the last five decades. Far from being a forensic anomaly, genetic genealogy is quickly on its way to becoming a routine police procedure. At least one company has begun offering a full-service genetic genealogy shop to law enforcement clients. And Rae-Venter’s skills are in such high demand that she’s started teaching her secrets to some of the biggest police forces in the US, including the Federal Bureau of Investigation.
Identifying individuals from their distant genetic relatives, a technique called long-range familial searching, is a potent alternative to the types of DNA searches commonly available to cops. Those are typically limited to forensic databases, which can only identify close kin—a sibling, parent, or child—and are highly regulated. No court order is required to mine GEDMatch’s open source trove of potential leads, which, unlike forensic databases, contains genetic bits of code that can be tied to health data and other personally identifiable information.
Currently, there aren’t any laws that regulate how law enforcement employs long-range familial searching, which hobbyists and do-gooders have turned to for years to find the biological families of adoptees. But some legal experts argue its use in criminal cases raises grave privacy concerns. They expect to see a legal challenge at some point, though probably not in the next year. In the meantime, GEDMatch is becoming even more powerful, as it grows by nearly a thousand new uploads every day. And with hundreds more cases currently in the hands of full-time family-tree builders, one thing’s for sure: In 2019, genealogy is going to send a lot more people to jail.
It was the last Saturday of June and CeCe Moore had been working on her couch, hunched over her laptop for 16 hours straight. The month before, the genetic genealogist had been hired by a forensic DNA company in Virginia called Parabon, to lead its new division devoted to long-range familial searching. She was immersed in a case out of Fort Wayne, Indiana; In the spring of 1998, eight-year-old April Tinsley went missing from her home. Three days later, a jogger discovered her body in a ditch on DeKalb County Road 68, about 20 miles outside of town. She had been raped and strangled to death.
For years, Tinsley’s killer haunted that northeastern corner of Indiana, leaving messages scrawled on a barn bragging of his crime. In 2004, four threatening notes appeared on bicycles owned by young girls that had been left in their yards. The notes, which were claimed to be written by the same person that killed Tinsley, were placed inside baggies alongside used condoms. The semen matched DNA found in Tinsley’s underwear.
This summer, Indiana investigators extracted DNA from the original crime scene and sent it to Parabon. There, the company reverse-engineered the information into a DNA data profile similar to what you would get back from consumer genetics companies like 23andMe or Ancestry. Then they uploaded it to GEDMatch and waited for a match. They got 12. Twelve relatives, ranging from fifth to third cousins.
So that’s where Moore started, that weekend in June. The cousins represented four different family trees containing thousands of people, all of which somehow had to tie into the Fort Wayne killer. The first thing she did was work backward in time to locate ancestors from whom the suspect and the 12 matches were both descended. Eventually she found four couples, born between 1809 and 1849. Once she had them, she could move forward in history, building out family trees of every generation until the present. She did this by tracking names and faces through census records, newspaper archives, school yearbooks, and social media.
By the time night fell over her home in San Diego, she had begun to close in on a single branch, into which the four genetic tributaries all ran. From there things moved quickly. As the clock ticked past midnight, she found the relatives that had struck out for Indiana. It didn’t take much longer to circle in on two brothers who lived in the area where Tinsley was murdered. Full siblings are as close as genetic genealogy can get. But Moore had a hunch. One brother struck her as a recluse; he had no wife or kids, he lived in a trailer, there were no pictures of him anywhere, and his family never mentioned him on Facebook.
Moore laid this all out for the Indiana investigators. A few days later they came back to her with a photo of one of the two brothers, with a hand-written note underneath. She gasped. “I thought it was him, but I wasn’t sure until I saw his writing,” Moore says. “It was the same as those notes and that barn.”
Indiana authorities staked out the trailer the first week of July and collected a piece of trash with the suspect’s DNA on it. Lab tests confirmed that the DNA recovered from the condoms in 2004, and the crime scene in 1988, belonged to the same man: 59-year-old John Dale Miller. Police arrested him July 15th. According to local reports, when the police asked him why they were at his home, Miller replied, “April Tinsley.” On Friday, December 7, Miller pled guilty in the Allen County Courthouse to murder and child molestation, as part of a plea agreement. On December 21, a judge sentenced him to 80 years in prison.
Miller is the first person genetic genealogy has put away for good. There could soon be others. Parabon has made public its involvement in 20 “solved” cases so far, with eight more undisclosed. Of those, at least four of the suspects were already deceased. The company was able to get up and running quickly after the Golden State Killer news broke the ice, because it had already created 100 or so genetic profiles through its phenotyping service, which creates a composite image from DNA for police to circulate in hopes of getting a lead. After hiring Moore, they quickly brought on three more genetic genealogists over the summer, and are in the process of hiring one more. The company says it’s now uploaded 200 profiles to GEDMatch, which represent cold cases from nearly as many law enforcement agencies scattered around the US. Parabon is actively working on 40 such cases.
Some of those include active cases, not just crimes committed decades ago. For example, in April, exactly a week before the Golden State Killer announcement, a man broke into a residence in the red-rocked southern Utah town of St. George, and sexually assaulted the 79-year-old woman who lived there. Three months later, authorities arrested a suspect, Spencer Glen Monnett, based on Moore’s genetic detective work. She says everyone clears their schedules to prioritize any active cases. At the moment, Parabon is working on at least one other active case involving a serial offender, but the company expects it to become a bigger part of what they do in 2019.
“In these active cases that come back with no match in CODIS [the federal offender database], law enforcement are realizing they don’t have to wait until every last avenue has been exhausted before coming to us,” says Ellen Greytak, who runs the company’s advanced DNA services division. “Genetic genealogy can be a tool to use right away.”
Rae-Venter, the genetic genealogist who cracked the Golden State Killer case, has also begun taking on active cases, with a small team of volunteers who mostly work pro bono. Right now she’s spending 12 to 15 hours a day, six days a week trying to track down a serial rapist who is still out there committing crimes. In addition to that one, her consultancy group is working through a backlog of 25 to 30 more cold cases. And she’s still collaborating closely with the Sacramento County detectives she joined forces with on the GSK case. Rae-Venter says a large part of their queue are referrals from the FBI.
Indeed, the feds can’t seem to get enough of her. Earlier this year, The FBI flew Rae-Venter to Houston, Texas to give a seven hour presentation on genetic genealogy to a room of 100 people—mostly federal agents, some local police officers, and even one Texas Ranger in a signature white rancher-style cowboy hat. “It’s really catching people’s attention,” she says. While family historians such as herself may be leading the way in this emerging field, she thinks it makes more sense to train, and perhaps even certify, law enforcement, rather than try to pull from the hobbyist community. Ultimately, she believes every major law enforcement agency will have its own specialists on staff. “I think this belongs to detectives, not genealogists,” says Rae-Venter.
As an example, she points to the arrest in September of the man believed to be the NorCal Rapist, another serial offender who terrorized victims in six California counties over a 15-year period beginning in 1991. Detectives from the Sacramento District Attorney’s office, who Rae-Venter had trained, uploaded a genetic profile of the suspect and built out family trees on their own. According to the DA’s office, they singled out the man arrested, Roy Charles Waller, in just 10 days.
Genetic genealogy alone isn’t enough to make an arrest. Investigators have to do confirmatory DNA testing, by retrieving bits of genetic material from the suspect, usually pulled from his or her trash, and comparing them to DNA found at the crime scene. But legal scholars worry that the widespread adoption of long-range familial searches will expose vast numbers of innocent people to genetic surveillance.
GEDMatch, which currently houses 1.2 million profiles from folks who’ve had their DNA analyzed at places like 23andMe and Ancestry, can now be used to identify at least 60 percent of all Americans with European Ancestry, regardless of whether they themselves have ever been tested. That’s according to two recent analyses by genetics researchers, who expect databases like GEDMatch to grow so big in the next few years that it will be possible to find anyone from just their DNA, even if they haven’t voluntarily put it in the public domain.
“You can’t claw back the profile of your third cousin once removed who you don’t even know exists,” says Erin Murphy, a law professor at New York University Law School and an expert on familial DNA searches. If someone gets ensnared in a long-range familial search, she says, they’re going to have very little legal recourse. “These searches throw into sharp relief how current privacy protections under the 4th Amendment are insufficient to contend with what technologies are available to police in 2018.”
There’s not a lot of data yet on whether the general public believes police should have access to non-criminal genetic databases. But initial surveys suggest that the majority of Americans are most supportive of such searches when they’re used to go after violent offenders. Approval drops from 80 percent to less than 40 percent for pursuing people who’ve committed nonviolent crimes.
As long as genetic genealogy remains resource-intensive, it’s unlikely to be used to catch thieves or drug-dealers. But even that appears to be changing. As more people send in their spit to DNA testing companies and upload their results to GEDMatch, matches will become more frequent. Curtis Rogers, the website’s 80-year-old co-creator, says the site has grown by about 200,000 profiles since it changed its terms in May. And its search capabilities just got more powerful. Last week, Rogers’ team of retired computer engineers rolled out an update that will let people see more matches with even more distant relatives. They also added a tool he calls “revolutionary” that will help sleuths decide if a small piece of DNA is a real match or just noise. Since the website’s skeletal staff are mostly volunteers, the effort has taken about two years. But now that it’s done, Rogers says they can start thinking about how to transform their hobby site, which resembles a Web 1.0 Wiki, into something more professional—including adding backups and beefing up security. “We want to make sure it’s something that is around for the long term,” he says.