Thanks to “genetic genealogy,” solving crimes with genomic databases is becoming mainstream — with some uncomfortable implications for the future of privacy.
Credit…Illustration by Delcan & Co.
On any given day, CeCe Moore’s inbox is flooded with strangers asking her to solve the mysteries of life and death, and on a good day, she can. Over the past year, working from anonymous DNA samples, Moore helped identify the suspect in a murder by tracking old migration patterns from Poland to northern New Jersey; solved an assault case in which female DNA at the crime scene turned out to have been left not by the usual sort of explicable coincidence but by an actual female perpetrator; and made a key discovery in a murder investigation when the gravestone of an apparently childless woman was found to bear the inscription “MOTHER.” Some of these she knocked out in a few frenzied days. But there was also one cold case that had dogged her for close to two and a half years: the murder, in Gresham, Ore., of a woman named Barbara Tucker.
Moore is perhaps the most prominent figure in the field of genetic genealogy — the mapping and measuring of how relatives share DNA — which was developed in the aughts as a result of direct-to-consumer genomic testing. One application of genetic genealogy, it quickly became clear, was to determine someone’s identity through a broader web of genetic relationships, and Moore rose to prominence solving maternity-ward mix-ups, helping adoptees and foundlings locate birth parents and identifying unknown fathers. Starting with relatives in DNA databases and working through public records and social media, you could narrow in on the right branch of the right family tree, triangulating the mystery person; in other words, you could use the haystack to find the needle.
For years, Moore had been curious and wary about the idea of applying her technique to criminal cases. Employing data from genomic tests toward that end struck her as off-label. Then, in April 2018, law-enforcement officials in California announced that genetic genealogy had been used to identify the serial rapist and murderer known as the Golden State Killer, a former police officer named Joseph James DeAngelo Jr. The story’s wide exposure — and almost universally positive reception — satisfied Moore’s qualms about user consent, and within a few days she was working her first homicide. Her identifications led to the first exoneration of an innocent man, the first trial and the first jury conviction resulting from genetic genealogy. So when “60 Minutes” featured her in a segment that October on cold-case breakthroughs, Moore braced for the familiar surge of requests, from the delusional to the unrefusable. One of the latter was from a woman named Susan Pater — nee Tucker — imploring Moore to take a look at her sister’s case.
Barbara Tucker was walking to her business class at Mt. Hood Community College one rainy evening in January 1980. At 19, she was the baby of seven siblings and half-siblings sharing the same mother, but stood out all the same — confident, outgoing, the tallest girl on her high school basketball team. Motorists later reported seeing her attempt to wave down cars, stepping in front of oncoming traffic, her face covered in dirt or blood, as a man approached her from the bushes. “Nobody stopped,” Detective Aaron Turnage, who inherited the case in 2015, told me. “It was a flash covered by the view of their headlights — it was over and they moved on.” Some thought they had happened upon a college-student prank. Tucker’s body, bearing the signs of beating and sexual assault, was found amid snow and shrubs at the edge of a parking lot the next morning.
Gresham’s Police Department subjected at least a dozen men to polygraphs. The clearest description of the suspect, the police told reporters at the time, came from a witness under hypnosis. (Turnage was skeptical that these reports were accurate.) The leads slowed, newspaper updates ran shorter and further apart, a $10,000 reward went unclaimed and by 1989 the case had gone cold. A semen sample had been preserved from the crime scene, so the emergence of criminal DNA profiling, in 1987, held out some hope, but in the years that followed no matches turned up in CODIS, the nation’s ever-expanding Combined DNA Index System, which is administered by the F.B.I.
By the time she took the case, Moore had teamed up with Parabon NanoLabs, a Virginia company whose services include novel DNA-based forensics. Now, with the go-ahead from the Gresham Police, Parabon’s director of bioinformatics, Ellen Greytak, uploaded a DNA profile of the semen sample to GEDmatch, an open-access site where users of testing services (the biggest two are 23andMe and Ancestry) can upload and compare profiles. The suspect and his nearest genetic relative in the database shared vanishingly little DNA, with their most recent common ancestor most likely several generations in the past.
After an initial evaluation, Moore gave the case the lowest score on her vast spreadsheet of investigations: 5 out of 5 — unlikely to be solved. But she continued to poke around, largely pro bono, stealing hours here and there, building hundreds of family trees and hoping for a break as she tried to pull one man out of countless possible candidates. She didn’t mention her progress, or lack of it, to Turnage or to Pater, not wanting to get hopes up. The one clue revealed by the suspect’s genome was the unusual phenotype of brown eyes combined with bright red hair.
Then, this past March, another distant relative turned up in GEDmatch. Greytak emailed Moore with the news, signing off, “Go get him, tiger.” Moore logged on and saw that the new profile might pull the puzzle pieces together. Within days she submitted her findings to Turnage: The suspect had to be among the seven grandsons of a single couple, and she noticed that one had been through an unstable childhood and had served time for kidnapping. This was Robert Plympton, a wilderness guide who led fishing trips down Oregon’s Sandy River. Now married with a son, Plympton was an athletically built 16-year-old at the time of Tucker’s death and lived less than two miles from where she was killed.
In May, Moore told me that a suspect in a case of hers should be expecting a house call from the police soon. I later found out she was referring to Plympton. She had already changed her spreadsheet entry from “completed” to “solved” after he spit out his chewing gum on a bike ride and his DNA matched that from the 41-year-old crime scene. If you check Plympton’s inmate profile at the Multnomah County jail, where he’s awaiting a March 2023 trial, you’ll find you can make out the red hair in his goatee.
Part of what pushed Moore forward in her work on the Tucker homicide was an age-old force in such cases: pressure from the family of the victim. But the truly decisive factor was not the involvement of Tucker’s close relatives; it was the involvement of Robert Plympton’s distant ones, who almost surely never met or even knew of him. They were the ones who had sent a tube of saliva or a cheek swab off to a private company for sequencing, then shared that data even more widely. They were the ones who — thanks to a few crucial shared spans across three billion pairs of A, C, G and T nucleotides — brought about his arrest.
The first sequencing of an entire human genome took more than a decade and cost around $3 billion, wrapping up in 2003. Within five years, a new generation of mass-produced, low-cost genotyping chips debuted that could perform an almost instantaneous high-resolution sampling of a genome. Soon, millions of people were generating profiles with hundreds of thousands of data points, of a kind called S.N.P.s (pronounced “snips”). (For comparison, the standard genetic profiling used by law enforcement since the 1980s consists of at most 20 genomic data points.) Consumers now might be able to see what Irish province they traced back to, and what genes made them cilantro-haters or twitchy sleepers. They could also see a menagerie of new relatives, enabling forms of familial research that scientists regarded as an inconsequential sideshow and left to amateur enthusiasts to tease out.
DNA was a messy addition to genealogy, a sleepy, prim discipline of aristocratic origins in which the paper trail was sacrosanct. Just mentioning genetics on a genealogy email list could get you banned. And even among those who embraced the new technologies — limited forms of consumer profiling had been available since 2000 — most remained interested primarily in the distant past, not the search for recent or living relatives. Moore was an exception. A decade ago, she helped found a group that figured out how to fuse DNA with traditional techniques for finding adoptees’ birth families. Then, in 2013, she established DNA Detectives, which brought together parentage seekers with those who could solve their cases or train them to do it themselves. The group now has more than 170,000 members on Facebook.
Some in the field fretted that such investigations could violate the anonymity of birth parents or egg and sperm donors, revealing painful family secrets (incest, rape and infidelity), all of which might turn people against sharing their data freely for tamer pursuits. Later, there were fresh objections to identifying John and Jane Does, which had the potential to entangle genetic genealogy, indirectly, with homicide investigations. John and Jane Does were also not living, willing participants, which meant that 23andMe and Ancestry were off limits — both companies accept only fresh DNA samples, not uploads. (The two companies have more than 32 million profiles between them.) But GEDmatch did, and once its database became big enough, something significant changed. A technique developed to identify someone who desperately wanted to know who he really was, after all, could also be used to identify someone who desperately didn’t want anyone to know who he was.
Throughout all this, genetic genealogy stayed a fairly cozy scene — a sandbox its practitioners had all built and played in together, everybody participating in everybody else’s Facebook groups, conference panels and seminars. But after the Golden State Killer arrest, the stakes in genetic-genealogy cases began to involve literal life or death. This exacerbated old tensions and caused new ones, making universal collegiality impossible, especially for Moore, who had been central to both the core genetic-genealogy scene and to the investigative vanguard. There was legitimate fear, she said, that law enforcement, using what genetic genealogists had collaboratively built, might “destroy it for the purpose for which it was created.”
The parochial squabble had also become a subject of far wider interest. Law-enforcement agents now had access to a “genetic panopticon,” to borrow a phrase from Justice Antonin Scalia, created not by the government but by the citizenry. The underlying debate often seemed less about what rules, laws or standards of consent ought to apply, and more like a grasping for concepts and analogies: Was this use of genetic informants creepy? Was it unfair? Was this a cautionary tale about clicking “I agree” when asked whether all kinds of our personal information could be used for purposes we often failed to understand? Had the myriad uploaders unilaterally altered some fundamental definition of privacy, or had they just shown us that a change, already long underway, could no longer be stopped?
The path that led Cynthia Moore to the pinnacle of genetic genealogy, in particular of the mystery-solving type, seems at once unlikely and inevitable. If you look at Mt. Carmel High’s 1986 yearbook on Ancestry (it’s a scan of Moore’s personal copy on the site), you’ll find an earnest musical-theater nerd. Moore matriculated at the University of Southern California the next year, in the vocal-performance program, switched majors and left college one final shy of a degree. She worked in musical theater and then spent more than a decade in the hinterlands of Hollywood: television commercials, industrial films, modeling and working at trade shows as what she jokingly called a “booth babe.” She played Barbie at so many events that a Mattel designer told her that the company had created a doll in her likeness, and to this day she has to fight muscle memory not to sign autographs “Barbie.”
During this period, Moore was haunted by a sense of unfulfilled ambition and a feeling that she was “wasting her brain.” She gravitated toward an obsessive mastery of systems: She used rudimentary internet databases to vet her friends’ suitors and taught herself the arcana of credit repair and the family court system (after a tough separation from her son’s father). She also spent a fair amount of time on true-crime forums, familiarizing herself with some of the cases she would later solve.
Moore had a few close calls herself. In high school, on her way home to San Diego from Los Angeles, where she was performing in a play, she was pulled over on a lonely stretch of Interstate 15 by a California Highway Patrol officer. The encounter escalated frighteningly: The officer dragged her from her car by her arm, berating her, and grew enraged when she cut her bare feet on the roadside gravel. She was left unnerved, but she didn’t report the incident. A few months later she learned that the same patrolman, Craig Peyer, had pulled over another young woman, then bludgeoned and strangled her to death. (Several years later, she earned a commendation from the C.H.P. when she pulled a burly, unconscious sheriff from a burning car.) During college, a trespassing incident baroquely escalated into a local street gang targeting her for rape and murder, requiring a police officer to escort her to campus.
Genetic genealogy provided the final system to master. Moore had been a keen hobbyist in traditional genealogy and in early consumer genomics, which had allowed the tracking of direct male lines through the Y-chromosome and of maternally inherited mitochondrial DNA. But with the revelatory debut of 23andMe’s Relative Finder tool, in 2009, she went all in. She took over the management of a major DNA email list, started answering user questions and began blogging. Her blog’s traffic logs soon showed visits from major research universities. When she invited Henry Louis Gates Jr., the literary scholar and host of the celebrity-genealogy PBS show “Finding Your Roots,” to deliver a keynote address at a genetics conference, he lingered to catch Moore’s own talk. Gates, who had made farsighted plans to start using S.N.P. profiles for his show, was so impressed that he offered her a job as his resident genetic genealogist. “She’s a pioneer,” he told me, “and she’s an autodidact. It’s mind-boggling.”
These days Moore lives in a spare modern house Jenga-blocked into coastal Orange County’s San Joaquin Hills with her 16-year-old son, Nicky, and her partner, Lennart Martinson, a Swedish-born Southern Californian of the Jeff Bridges phenotype. She keeps to a night owl’s schedule, even when she rises early for a status update with detectives on the East Coast. Long before it became commonplace, her days demanded a numbing succession of video calls with full makeup and PJ bottoms. If she didn’t have a call, or if I wasn’t visiting, she might work from bed.
I had seen Moore in frantic motion out in the world, at news conferences for arrests and shooting back to back for daytime talk shows or TV documentaries or her ABC series, “The Genetic Detective”; this was her other mode: settling into weeks or months of sedentary monomania. The house had an atmosphere of monastic discipline, with Martinson at the kitchen table, editing footage in Final Cut for his production business, Nicky disappearing for tennis practice or naps of heroic length, and Moore curled up on an enormous, well-worn sofa in the living room, using an equally antiquated MacBook with far too many tabs open. While working on cases, Moore occasionally punctuates the long silences by responding to ambient TV (talking back to a show that covers a case of hers without mentioning who identified the suspect) or reading her email out loud, including the spam (“What does this mean, ‘John Goodman confirms the rumors’?”). Several nights in a row, as the lingering Pacific twilight faded outside the picture windows, I realized nobody was going to turn on the lights and got up to do so.
Moore still receives requests from the offspring of incestuous unions; the people whose genes indicate that they should never have been viable embryos; the petitioners who, harboring something less than legitimate suspicion, ask her to explore whether they were stolen at birth. “It’s never true,” she says. “They’re always just the children of the parents that they hate.” (Moore did take a few days off earlier this year to prove that her own aunt’s child was stolen and sold as a newborn.) But aside from her ongoing duties on “Finding Your Roots,” these other mysteries of human connection have been eclipsed by her investigative caseload. That fervent commitment stems, she says, from the same beliefs that guided her hard line on adoption work: a right to identity, knowledge and resolution that overrides other concerns.
Moore is also clearly driven by the high-adrenaline, compulsion-making nature of the work, in which puzzle-solving is improbably raised to a level of extreme consequence — as if every time you completed a crossword, you incidentally prevented a car crash. I got a firsthand sense of this when Moore, amused by my presence in her house, decided to put me to work. She’d had a hard time uncovering the maiden name of a woman in a Jane Doe case she was working on. I had an instinct to check property records, which sometimes reveal personal information in roundabout ways, and found a joint-tenancy affidavit that the woman had filed with the county when her husband died intestate; buried in it was a copy of his death certificate, which otherwise would have been unavailable to the public and which contained her maiden name. I immediately asked Moore for other tasks, and back at my Airbnb, as the sky grew light outside, I fell asleep with my laptop open to FamilySearch.org.
Investigative genetic genealogy moves backward before it moves forward. That is, starting with the target’s genetic matches, you trace the matches’ ancestors and then those ancestors’ descendants, finding any points where the matches’ lines intersect with one another, closing in, ultimately, on possible candidates.
Moore walked me through the case of Jay Cook and Tanya Van Cuylenborg, a young Canadian couple who were murdered in 1987 while visiting Seattle — the first criminal case she worked and also the most straightforward. (The defendant’s June 2019 conviction, the first for a defendant identified through genetic genealogy, was vacated in December because of juror bias.)
The initial step in any search is to generate the target’s S.N.P. profile from the available physical evidence, upload it to the database of choice (in this case, GEDmatch) and try to put names to the other profiles, or “kits,” with overlapping DNA. The average person has an orbit of roughly 850 genetically significant relatives, so even with a very small number of European Americans in a database, you can cover a huge swath of the United States population. (Moore estimates that about 500,000 kits are currently available to law enforcement on GEDmatch.) The amount of shared DNA tells you roughly how far back to build — whether you’re looking for a common ancestor at the grandparent level or the great-great-great-great-great-grandparent level.
A single match is basically unusable (unless genetically close to the suspect); two matches who are unrelated to each other allow for the necessary triangulation; a better, yet more complex, scenario, Moore told me, is to have several sets of matches who are related to one another but not to the other sets of matches — only to the target. Then all the relationships act as crosschecks.
On GEDmatch, each kit has a name and an email address. In the Seattle case, the suspect DNA had two decent matches, neither of which was related to the other. One was Chelsea Rustad, using her real name and personal Gmail address. The other — let’s call him Little Y — was harder to ID. His kit gave only initials, and those didn’t line up with the email. Running the address through Google and people-finder databases pointed to a woman who turned out to have a grandson with the right initials: Little Y seemed to be a teenager living in the American Southwest. That was enough for Moore to keep moving.
Then comes the building backward and forward, a process that draws in hundreds of relatives who will ultimately prove irrelevant. Chelsea’s father, Mark, was born to Ormond Rustad and his wife, Colette; and Ormond, in turn, was the son of Ole Rustad and Janna Sund — two of Chelsea’s great-grandparents. On the other match’s side, Little Y was born to Daddy Y, who was born to Grandpa Y, who was the son of a Mr. and Mrs. Talbott. And Mr. and Mrs. Talbott had another son aside from Grandpa Y: William Earl Talbott Sr. Ole and Janna Sund, meanwhile, also had a daughter named Blanche, who married Edwon Peters. Blanche and Edwon’s daughter, Patricia Peters, was married to William Earl Talbott Sr. The trees of the two unrelated matches came together in the union of William and Patricia, and they had only one known son, born in 1963.
On May 8, 2018, that son, William Earl Talbott II, pulled up to the intersection of West Marginal Way and Spokane Street in Seattle as police followed him. He momentarily opened the door of his semi truck to reach around behind the cab, and let fall a coffee cup with DNA that matched the semen found in 1987 on the leg of Van Cuylenborg’s trousers.
It took Moore two hours to find Talbott. Her quick and dirty process might be described as brute-force cross-referencing reined in by intuition. Rather than commit everything to notes, she leans heavily on a capacious working memory. The search leads to newspaper archives major and minor, obituaries and wedding announcements, high school yearbooks and college alumni updates, gravestones and official records — of marriage, divorce, birth and death, whose availability depends on a legal and technological patchwork of federal, state and local sources, some of which are well digitized and centralized, some not. And of course, the all-knowing, all-oversharing fount of social media, which demands its own form of minute analysis.
A frequent challenge in searches is to push a match’s family history back through the most recent available United States Census; these are made public after 72 years, meaning 1950 will become available in April. “The hardest thing can be to find someone’s grandparents,” Moore said. “Once you get back through that bottleneck, the records open up.” Even once clear of the census bottleneck, the backward search can unfurl for centuries in the world of ship manifests and Old Country marriage registers and requires a ceaseless struggle with information that is ambiguous, incomplete, wrong or missing. Moore might, for example, run out of years to tie together the strands of an African American family before coming up against the informational void of slavery, and even if she can grope through that, with tax documents listing human chattel or census rolls that included free people of color, maybe when she tries to track their descendants she’ll find they all chose different surnames then and scattered in the Great Migration.
Once the DNA has told you that you’re far back enough to start building forward, you might get lucky and quickly find where two of the lines cross; or you might have to look far afield, as when Moore traced a bachelor from one line to a small town in the Pacific Northwest where, for a brief period in the 1930s, he lived suggestively near a married woman from the other line. As Moore starts filling in descendants in different trees, something may ping — a surname spotted elsewhere, the city in which the crime took place, the unusual ethnic mix of the suspect sample — that shows at least one worthwhile path. It’s easier for things to go wrong with descendancy, for a simple reason: Everyone has two biological parents, but you never really know how many children to look for. If somehow you miss a child, you could miss the line that leads to the suspect.
At every stage, Moore must gauge relatedness in the paper trail against relatedness suggested by the actual DNA and resolve any discrepancies that arise. The genetic distance between the kits has to make sense or it all falls apart.
The 23 chromosomes are supposed to come in pairs, which means you get one from each parent via sperm or egg — and for that to happen, each of their 23 pairs has to randomly recombine into a single, new version, losing half the original material in the process (though the X and Y chromosomes for males don’t recombine in this manner). Much of the physical variation we see within and between generations is a result of this halving of genetic material. You should inherit half of your autosomal DNA — the genetic material from the 22 non-sex-determining chromosomes — from one parent. The length of those shared identical autosomal segments will total about 3,600 centiMorgans, a unit for measuring spans of DNA, while you and a first cousin or great-grandparent, at two additional degrees of remove, should share about a quarter of that.
It’s these fungible quantities of coding or noncoding DNA, not specific shared genetic traits, that Moore is looking at. They are what allow her to infer an unknown relationship or verify a known one. The further you are from a genetic relative, the shorter the shared identical segments become, until many of them vanish. So as the number falls, the relationships become less predictable; when Robert Plympton’s original top matches were all in the 20-30 cM range, Moore says, it was difficult to tell if they were tenth cousins or third cousins, making them unpromising leads.
Talbott and Chelsea Rustad, with 232.1 cM of DNA in common, were most likely “second cousins,” a shorthand that can describe other relationships at five degrees of separation. As it turned out, they are true second cousins, sharing a set of great-grandparents: The quantity of identical DNA correctly predicted how far back to build and also confirmed the paper-trail relationship. Little Y and Talbott are also “second cousins” by their shared DNA, 302.3 cM. But Little Y’s great-grandparents, Mr. and Mrs. Talbott, were William Earl Talbott II’s grandparents. Little Y and Talbott are first cousins once removed — and that relationship is a degree closer than second cousins, so they should share roughly twice as much DNA as they do. How did the DNA get cut in half? Outlier factors can sometimes affect sharing, but those seemed unlikely here; the answer, Moore found, was that William Earl Talbott Sr. was not, in fact, Mr. and Mrs. Talbott’s son — he was only Mrs. Talbott’s, by a previous marriage, and was adopted by his stepfather. His name changed when he was a toddler.
People who have questions about their parentage flock to consumer genetic databases and wind up overrepresented, so the matches in any search are disproportionately likely to turn out to have something wrong with their trees, if they’re not in the wrong forest entirely. And it’s not just them — parentage puzzles can happen anywhere up and down the chains of family relations, buried decades or centuries in the past. Just incidentally, Moore might have to figure out an entire extra genetic genealogy case, or two, nested inside the one she’s trying to solve.
Anytime the lines come together deep in the past, the genetic and geographic math may not narrow the descendants sufficiently. Then the process may turn to target testing: creating new, voluntary matches in the tree. Law enforcement may ask select relatives to provide their DNA or upload existing profiles to GEDmatch for an unspecified investigation. On occasion, Moore’s team might contact the relatives, and her visibility on “Finding Your Roots” can help persuade them that it’s not a scam.
The final step requires an ability to distinguish hunches from confirmation bias. For example, that a 45-year-old woman in 1920 with a teenage daughter and a newborn was probably the baby’s grandmother, or that an old East Coast line will connect to a suspect if you can spot what Moore calls the “adventurous DNA” that struck out West a hundred years ago. Even now Moore is still sometimes touched by the strange intimacy of getting to know families in ways the members themselves might not: the lovers shot in the street and great-grandchildren who meet the same fate, the chaos of intergenerational poverty and privation, the orphans and vanished parents, the infidelity and infertility.
Moore says Talbott was a “ghost” who left little trace, and she had sensed he kept to himself. I asked why. An obituary for his mother had specified his three sisters each as being “of” a specific town, she said, and then there was just “ ‘Bill Talbott in Washington.’ I wondered if he was even living a normal life at all. Why does he get a different preposition?” She later heard that Talbott had drifted away from the family after he broke his sister’s arm. Moore noticed that the only relative at the sentencing hearing was Rustad, who was there to support the victims’ families.
Two and a half years ago, in May, Moore told me that a big case was about to break. She couldn’t yet reveal what or precisely where, but I should fly to Salt Lake City, the closest major airport, and await further instructions. After I booked my ticket, I sketched out Salt Lake City’s hub zone by excluding the radii of Denver, Phoenix and Vegas, then went to Parabon’s website. Since 2014, the company has offered a service called Snapshot that produces phenotype mug shots of unknown suspects, predicting hair and eye and skin color as well as facial structure. Because these are based on S.N.P. profiles that can also be uploaded to GEDmatch, Parabon had wound up with dozens of potential cases on hand when Moore decided to dive into investigations. There was one public Snapshot case that fit the geography, in Idaho Falls. After I landed, and just before midnight, Moore forwarded me the announcement of a news conference the following day.
That morning at the Idaho Falls Police Department was a slow one, which is to say an average one given the size of the town, punctuated with residents wandering in to ask about a missed jury summons or surrender old medication. And then, in the space of about half an hour, dozens of people turned up. There were tweens and great-grandmothers, toddlers and adults just to the far side of middle age, a number of the women bearing a resemblance to one another and also — it was clear if you had studied the posters tacked up in the hallway — to “Angie Dodge, homicide victim, 6/13/1996.”
“There’s not a person in this city who doesn’t know the story of this case,” the police chief, Bryce Johnson, told me. “It has defined the police department. It’s part of the DNA of it.” Eventually, the crowd headed down the street to watch Johnson take the lectern in the City Council’s chambers, where he told the room that, the previous afternoon, near the Oregon border, officers had walked up to a 53-year-old man named Brian Leigh Dripps Sr., asked him to come in to the local police station for a chat and coaxed forth a confession. Then Moore stepped up to explain the long research process, detailing how Dripps — the missing seventh direct male descendant of the right couple — had evaded her; when she finally found him, she realized he had been Dodge’s neighbor in 1996.
This was my first time seeing Moore, and she seemed to conjure, with her long double-helix curls and black pantsuit and sleek rectangular glasses, a glamorous yet relatable detective in a TV procedural. She spoke in crisp paragraphs, with a preternatural ability to hit all her marks. (Moore has racked up dozens of IMDb credits for playing herself.) After the conference ended, she navigated effortlessly among disparate parties, including the police, judicial-reform activists, journalists and Angie Dodge’s mother, Carol, with whom she had been in touch throughout her investigation. In one episode of “The Genetic Detective,” Moore drives around Idaho Falls with Carol and works her way through Dripps’s superimposed family tree like Tom Cruise in “Minority Report.” With every victorious arrest announced, with every prime-time mention, investigative genetic genealogy was finding new audiences, wider acceptance and more opt-ins. The victim profiles were also becoming more varied, no longer just white, middle-class and female. Moore evangelized not by argument but by loving the work and looking hypercompetent while doing it.
“What happened?” Karole Honas, the plain-spoken doyenne of the local ABC affiliate, asked the police during the Q. and A. She meant that Dripps had no other serious legal trouble. “He just went wacko bananas?” Later, outside the police department, amid Mylar fund-raising pinwheels and the last falling crab-apple petals, Moore told me that, yes, in her best guess Honas was right: In the days before the murder, Dripps’s first child was born and his wife had been trying to leave him, and it seemed as if something had simply gone wrong in his head. Many of her cases resemble this, with offenders who commit one horrific offense and then largely stay in line, though that analysis rests on many assumptions. Only a minority of violent crimes leave behind pertinent DNA evidence, a problem that may be compounded by broad awareness of forensic science and how it can and cannot be evaded. Police noticed that one suspect identified through genetic genealogy had a three-month-old newspaper on a table, open to an article about Talbott’s arrest.
On a day like that one in Idaho Falls, Moore seemed like a magician who could pull off any trick. In her final flourish, she had also cleared the name of Chris Tapp, who had served two decades for Angie’s murder despite having no DNA link to the crime. At the news conference, Tapp was profuse with hugs, handshakes and tears, and in another two months, he would be formally exonerated by a judge. “It’s the only unqualified joy I’ve experienced in law-enforcement work,” Moore told me. “Everything else has been so — heavy.”
Two days later, GEDmatch became all but useless to Moore.
Following the Golden State Killer arrest, in 2018, the site had posted a warning to users that police were uploading profiles, and hastily instituted a policy restricting such use to homicides, sexual assaults and unidentified bodies. But a few weeks before the Idaho Falls announcement, it emerged that one of the site’s founder-operators had, in a somewhat naïve, grandfatherly way, made an exception for a detective in Utah investigating a recent attempted murder. Moore was the one tasked with identifying the suspect (and did). Around the same time, it also emerged that FamilyTreeDNA, a consumer site with more than two million users, had been discreetly allowing the F.B.I. to upload suspect profiles to its database for genetic-genealogy searches.
GEDmatch scrambled to opt all accounts out of law-enforcement searches by default. Overnight, Moore’s available matches went from over a million profiles to zero, and her ability to work new cases practically vanished. “People will die,” she told CNN. In the months that followed, the handful of genetic genealogists whom she had recruited to build out the Parabon team had their hours cut, and she spent most of her time toiling on old cases for which she already had the list of matches.
Whereas some of Moore’s takes on the controversy were little more than rhetorical barbs, others were a reasoned apologia for genetic genealogy: The wrong people were not being implicated, and if innocent people were being entangled, the same thing happened in routine police work. The Utah attempted murder notwithstanding, there was little mission creep toward trivial crimes. And the wrongfully convicted, like Chris Tapp, now had a chance to overcome the inertia of the appeals process. Moore frequently emphasized that she and her team played by the rules and respected whatever safeguards were in place.
As FamilyTreeDNA refined its own search policies and GEDmatch’s opt-in numbers slowly rose, Moore started solving fresh cases and seemed outwardly unscathed. But the new database policies hadn’t actually resolved much. Some government investigators apparently just ignored them. Strictly speaking, law enforcement is entitled to see the same things any member of the public can, while also being freer to disregard the terms of service, so in some cases, they and their genealogists uploaded to GEDmatch without declaring themselves or used MyHeritage, a consumer site larger than GEDmatch and FamilyTreeDNA combined, which officially prohibits law enforcement use. (A motivated investigator who wanted to infiltrate 23andMe or Ancestry might conceivably be able to finagle crime-scene DNA into a saliva kit.)
Most of the consent debate had overestimated the importance of the “rules” by which law enforcement was asked to play. It had also incorrectly assumed that at stake was a straightforward concept of privacy — one person, making an individual choice about his or her exposure.
“There’s something that was really eye-opening to me,” Moore told me one day from her sunny couch perch. In 2012, she attended a genetics conference where she sat on a panel that included two women who were members of Native American nations. “The women explained they wouldn’t take a test without consulting everybody else in the tribe, because they’d be making the decision for everybody.” The same logic, she pointed out, now applied to the whole country. “It all happened under the radar, and it doesn’t really matter if you’re opposed: It’s a collective decision that’s already been made. A lot of what the privacy advocates have said I agree with. But 30 million people made that choice for everybody else.”
Moore’s candid assessment cuts to the heart of the unease over genetic genealogy and neatly describes a problem for which tech anthropologists have coined the term “networked privacy.” This really means lack of privacy: how you are exposed by others, how your individual choices can be diminished or undone when others can share things about you that you might not share about yourself. And so the conundrum around the expanded use of genetic databases is philosophical as much as legal: It’s about how we can control something that is uniquely ours — and yet not entirely ours to control. It’s about data that is personal yet transpersonal. As with the prime object lesson in networked privacy — social media — so with consumer DNA sites, the real breach is simply that the data exists in the first place.
Even the idea of a “right to privacy” is neither very old nor necessarily obvious. It has been called a penumbral right: It isn’t explicitly encoded in the Constitution or the Bill of Rights, and its evolving definition relies on other, more basic concepts, such as freedom of conscience and security of ownership. Louis Brandeis, as a young lawyer, laid the groundwork for our modern concept of privacy, arguing that “the term ‘property’ has grown to comprise every form of possession — intangible, as well as tangible.” The intervening 130 years of United States law have involved repeated clashes over when we ought to expect privacy and how technology changes those expectations. “The progress of science in furnishing the government with means of espionage,” Brandeis wrote in a Supreme Court dissent decades later, would never stop. What if it became possible, he wondered, to read a man’s very thoughts?
Technology has complicated the expectation of privacy perhaps nowhere more so than when it comes to consumers’ routinely sharing personal data with companies, which has itself undergone an epochal acceleration. As for your Fourth Amendment protections, sharing your information with a third party sometimes compromises them and sometimes doesn’t, in ways that can be weirdly specific and unintuitive. The thorny questions of third-party sharing don’t neatly apply to the criminal suspects in genetic-genealogy cases, because they haven’t themselves uploaded their DNA. This complicates the ambitious proposals by some law professors to protect the DNA you share with your cousin from warrantless searches in the same way that the cell tower pings you helplessly share with your mobile provider are protected. (There has been one instance where a suspect’s own profile was already in FamilyTreeDNA; it was created by the man’s daughter and granddaughter, but he died in jail before the issue emerged in court.)
New regulations and laws have begun to emerge, haltingly, to prevent encroachments and establish best practices. These are designed primarily to protect consumers, not suspects. (And once investigators in most jurisdictions take an interest in a suspect, there is nothing to stop them from surreptitiously collecting a confirmatory DNA sample.) The Department of Justice issued guidelines for federal investigators with some wiggle room, as well as, more bindingly, for any local agency receiving funding. Maryland and then Montana passed laws that established protocols and strictures for genetic genealogy’s use, the former with input from stakeholders including Moore. Canada, Australia, Sweden and other countries are in various phases of use or approval.
But genetic-genealogy investigations, still largely unconstrained, are here to stay. As Rockne Harmon, a former prosecutor and DNA consultant, told me, genetic genealogy is a nimble, privatized solution — “the lightning in the bottle,” he called it — that can leap over slow-moving bureaucracies and regulatory hurdles and, in some cases, over law enforcement’s own inefficiencies and failures. Debbie Kennett, a veteran genetic genealogist, has flagged a number of suspects identified through genetic genealogy who should have been in the F.B.I.’s CODIS database for prior or later offenses but apparently weren’t tested. One man who seems to have slipped through the cracks when he was in prison in the 1990s, according to Detective Turnage, was Robert Plympton.
DNA is the oldest network that exists among us, older than Facebook or marriage records, older than society or family, immutable and ineradicable. Yet, compared with the “social” networks that entice us to generate ever more data, to feed the omnivorous inhuman intelligence which seeks to know as many of us in as many ways as possible, genetic genealogy seems almost artisanal. It requires a discrete human intelligence working with single-minded purpose through trial and error and can succeed with a trivially small number of starting data points. It toys with the iceberg-tips of vast data regimes in unaccountable private databases, but it is nonalgorithmic, done by hand in pursuit of one right answer.
One way to think about the future of transpersonal data comes from danah boyd, one of the anthropologists who first described networked privacy: Just as choice is not really individual in the network, harm isn’t individual, either — it happens to everyone. A cultural shift can occur from the acceptance of many individual harms, which may be legal and legitimate, until the world in which we live is unrecognizably transformed.
Here is another way: The futurist Roy Amara offered a spin on the unintended consequences of new technology, known as Amara’s Law: “We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.” The pattern can be seen over and over: An innovation comes along; expectations run wild, then fail to materialize; but the new technology quietly expands, in unexpected ways, until it winds up having a profound effect after all.
Genetic genealogy was not the new technology — that, of course, was genomic mapping and its democratization, a cheap, fun tool to discover family and disclaimer-heavy information about disease risks. Rather, genetic genealogy seems to be a lesson or a warning. The problem isn’t whether the slope is slippery; it’s that we can’t see all the slopes we are on.
At the conference with the Native American panelists in 2012, Moore was taken aback by their reluctance to test without communitywide approval. “At the time it seemed so surprising, ‘What a unique way of looking at this!’” she said. But the intervening years had changed her view. “They were just ahead of their time. Now, it’s done.”
Rafil Kroll-Zaidi is a contributing editor for Harper’s Magazine, where he writes the monthly Findings column.