One day last July, two police officers paid an unexpected visit to the home of John D. Miller, 59, in the town of Grabill, Indiana. Finding Miller outside his house, where he lived alone, they asked him to accompany them. Miller, gray-haired and with a slight weight problem, didn’t resist. Not until they arrived at the police station, about a half-hour’s drive away, did the officers ask him if he knew what they wanted to talk to him about. Miller, the police report states, looked at the detectives and said, “April Tinsley.”
He was right.
Tinsley, a blond, smiling girl of 8, disappeared in the spring of 1988. She left the home of a girlfriend and vanished. Police were called in immediately, and three days later, a passerby came upon her body lying in a ditch about 20 miles from her home. She had been sexually assaulted and then strangled. Traces of the suspect’s DNA were found on the girl’s underwear, but the killer was never caught.
The case of April Tinsley became a cause célèbre in the United States. Off and on, over three decades, a team made up of personnel from the FBI and other federal agencies checked tips and examined numerous scenarios. The popular television program “America’s Most Wanted” devoted an episode to the murder in 2009, and in the internet era bloggers reported on developments in the case.
And developments there were. About two years after April’s death, police were called to a barn not far from where Tinsley’s body was found. Scrawled on the door was a message: “I kill 8 year old April Marie Tisley… ha ha I will kill agin [sic].”
- Genetic testing says Paul Ryan is 3% Jewish – what the hell does that mean?
- People have a right to disappear, France said. Then came a serial killer
- New DNA evidence found in 2006 murder of Israeli schoolgirl
In 2004, used condoms were found hanging from three separate girls’ bicycles in Fort Wayne, not far from Grabill, along with ominous notes: “Hi honey… I been watching you… I am the same person that kidnapped an rape an kill april tinsley... You are my next victim.”
But the killer didn’t strike again. In fact, that was the last time he was heard from. The investigation hit a dead end.
But then, early last May, 30 years after the murder, Virginia-based Parabon NanoLabs, which was working in tandem with CeCe Moore, a genetic genealogist, approached the Fort Wayne police, offering their services in investigating the case. Within a few days, they received authorization from the police to upload a DNA sample from the murder scene to GEDmatch, a genealogical site created by two enthusiasts from Florida and Texas. From there things moved quickly. All in all, Moore said, “It was rough. Afterward I was just totally beat. But I was determined to find him.”
Over the years, the police had narrowed down the list of possible suspects to about 100, but Miller, a recluse, wasn’t one of them. But Moore’s methods don’t require interviews with neighbors or potential witnesses. She compares DNA profiles uploaded to the web and reduces the number of potential suspects with the aid of other, open databases. It’s painstaking work but it can produce results relatively quickly – certainly when compared to three decades of fruitless searches.
Within 50 days, Moore gave police the names of two brothers. One of them, she said, speaking with Haaretz by phone, was their man. “It was challenge after challenge,” she said. “When the threads finally converged in Indiana, I knew I had got him.”
An arrest warrant requires evidence that’s admissible in court, so detectives scoured Miller’s trash for DNA samples and found what they were looking for in three used condoms.
At the station, the police told Miller that they had two DNA samples – one that was his, the other from the crime scene. And they match, he was told. Miller didn’t argue. That same day he confessed that he had kidnapped April Tinsley, raped her in his trailer and then strangled her to death so she wouldn’t tell anyone. He said it took him about 10 minutes to kill the girl.
“It’s amazing,” Detective Brian Martin, who heads the cold cases unit of the Fort Wayne police department, told me, in a phone interview. “We were able to do in less than 60 days what we were not able to do in 30 years. He was out there living in our community and we had no idea about him.”
In the past year, genealogical search engines have been used to solve complicated murder and rape cases, ones that had stumped police in the United States for decades. Vast sums had been spent on investigations, generations of detectives toiled on the cases, and the victims’ families lost hope.
Until CeCe Moore entered the picture. Moore, who was born in 1969 and previously performed on stage as an actor and singer, cracked the overwhelming majority of cold cases that were solved – within six months, and usually without even leaving home. All she needs is DNA from the crime scene, access to the GEDmatch site – which allows people who are searching for relations to upload DNA profiles at their own initiative – and a fast-internet connection.
The secret to unraveling the April Tinsley case, and many others like it, lies in the integration of tools from the realms of both genealogy and genetics. Genealogical research is a form of historical detective work that’s based on population records; media reports; marriage, birth and death certificates; archival research and so on. But recent developments in DNA-based technology have expanded the horizons of possibilities: Labs can now pinpoint information about familial origins and links for people even if they lack access to conventional documentation and records.
Moore’s gateway to the world of genealogy was a family tree she decided to create as a wedding present to a niece. Subsequently she became a well-known figure in the world of genealogy. She has helped draft ethical rules for proper usage of data, has often appeared on television to explain what she does, and she manages a Facebook group of 100,000 people, by means of which she teaches followers how to become “genetic detectives” and locate previously unknown relatives.
“I mostly worked with unknown parentage cases in the last few years,” she says, “helping people find their biological families when they don’t know who they are.” However, the challenge in such efforts eventually wore off. At some point, Moore came up with the idea using GEDmatch to try to solve cases of unidentified crime victims. That prepared her for the moment when she began to help law enforcement agencies, in partnership with Parabon NanoLabs, which specializes in DNA technology and had already been hired to help such authorities.
The results were not long in coming. Last June, police in Pennsylvania arrested Raymond Rowe for the rape and murder of Christy Mirack, a schoolteacher, in 1992. That same month, Gary Hartman was arrested in connection with the rape and murder of 12-year-old Michella Welch in Tacoma, Washington, in 1986. A month later, police in Utah used GEDmatch in investigating a more recent case: Spencer Monnett was arrested for the rape of a 79-year-old woman 14 weeks earlier. And the so-called “Ramsey Street rapist,” who had perpetrated serial assaults in Fayetteville, North Carolina, from 2006 to 2008, was identified as Darold Wayne Bowden, a white man with a large swastika tattooed on his chest.
All told, 19 cases were solved in this way, 17 of them by Parabon NanoLabs and CeCe Moore, the other two by different detective teams. Not one of the detainees had been considered a suspect.
One of the most complex cases Moore dealt with was that of the double murder of Jay Cook and Tanya Van Cuylenborg. The young couple from Canada – she was 18, he was 20 – was murdered in November 1987. The two were on their on their way to Seattle from Saanich, British Columbia, in the Cook family’s 10-year-old Ford Club van. They set out on November 18, and the last time they were seen was when they bought tickets for the ferry to Seattle, a little after 10 P.M. that day.
Five days later, Van Cuylenborg’s body was found in a wooded area in Skagit County, Washington; she had been raped and shot in the head. The next day, the empty van was found, and a day later all hope was lost when Cook’s body was located in a neighboring county, wrapped in a blue blanket. Jim Scharf, the detective who worked on the case in Snohomish County, Washington, said at the time of the arrest that over the years the police had received tips in connection with more than 350 people who might have been involved in the murders, but that none had shed any light on the events.
It took Moore just three days to give the police the suspect’s name.
Moore uploaded a DNA sample from one of the crime scenes to the genealogical search engine of GEDmatch. She was able to identify two distant relatives of the murderer, and with the use of traditional genealogical methods traced ancestors who were common to the suspect, and to every one of the relatives who had been found. At this stage, she started to build a family tree forward in time – in other words, she began to identify all the descendants of the ancestors right up to the present day and to clarify the relations between them – in a process known as “reverse genealogy.” The aim is to locate a descendant of the ancestor who matches the presumed suspect’s profile.
“The significant matches were at about the second-cousin level,” says Moore, adding, “I built family trees of people who were sharing DNA with the perpetrator, back to great-grandparents.”
It’s here that the significant difference from previous investigative methods lies, and where ethical problems arise. The search is not for the actual criminal – it’s unlikely, for example, that a serial rapist will upload his DNA to an open database – but for people who may have a certain genetic relationship to him but who may not even be aware of those ties and/or the suspicions against him. Moore has perfected this technique, and her recent successes have kept her very busy. Two months were needed to arrange an interview with her, but in the end she explained her work method to me, in a phone interview.
The starting point – the degree of the connection between the crime-scene DNA and the match on GEDmatch – is critical. This is because GEDmatch does not reveal the exact family relationship between the target of the search and the database, but offers a range of possibilities. “For any amount of shared DNA, there are multiple numbers of possible relationships,” Moore notes. “It could be a second cousin. It could be a second cousin once removed, a first cousin twice removed,” and so on.
Accordingly, Moore must first make an intelligent prediction of the nature of the connection between the suspect’s DNA and the results from the website. Most GEDmatch users upload their genetic information together with their full name and email address. Thus, when Moore identifies the correspondence bearing the strongest family relationship to the murderer or rapist, she has something substantial to work with.
In the case of the Canadian couple, Moore related that the family trees of two of the closest results did not share DNA. In other words, she knew that the place where the two trees overlapped would point to the killer. In fact, she discovered that the two trees converged in a marriage that produced one son. At that moment she had cracked the case: Last May, a 55-year-old man, William Earl Talbott II, was arrested for the two murders.
Talbott denied the charges, but the police are certain they have the right man. He was 24 in 1987, apparently lived about 35 kilometers from Seattle, and fresh DNA samples taken from him match the DNA found at the scene.
“If it hadn’t been for genetic genealogy, we wouldn’t have been standing here today, and if it hadn’t been allowed to be used in law enforcement, we would have never have been able to solve this case,” Detective Scharf stated in a press conference called to announce the arrest.
The reason that all these suspects were caught with the aid of GEDmatch and not through the FBI’s DNA database, known as CODIS, is related to the technology underlying each. CODIS has 17.6 million profiles, far more than the one million of GEDmatch, but the FBI’s goal is to compare samples and find a full match between them. This is accomplished by comparing short sequences called STRs (short tandem repeat) at about 20 locations along the genome. What it can’t do is identify with high probability ties between family members who are not first-degree relatives. The commercial companies, and thus also GEDmatch, use tests that examine genetic variants called SNPs (single-nucleotide polymorphisms) at 600,000 to 700,000 points on the genome. This technology makes it possible to detect distant family ties, and it’s precisely here that the sleuthing of amateur genealogists intersect with Moore’s skills and the world of criminal identification.
According to Dr. Steven Armentrout, the founder and CEO of Parabon NanoLabs, it’s difficult to calculate accurately the size of the net that GEDmatch casts. However, he says, 55 percent of the DNA profiles that Parabon has uploaded to GEDmatch since May have to date produced a lead sufficient to work with. His company, he notes, cooperates with law enforcement authorities across the United States and has so far uploaded about 200 DNA profiles.
Moore was not the first to use GEDmatch to solve murder cases. Its use began last April, with the arrest of Joseph James DeAngelo – the so-called “Golden State killer,” suspected of murdering 13 people and raping more than 50 women in California between 1974 and 1986.
Paul Holes, now a retired detective from the office of the Contra Costa County’s district attorney, spent 24 years pursuing the serial killer and rapist. When he was only months away from retirement, he decided to upload the DNA profile of the killer to GEDmatch.
What happened next could have been taken from a true-crime movie. The day before he was slated to turn in his badge and pistol, Holes found himself outside the home of a person named Joseph DeAngelo, wondering if he should simply knock on the door. He needed a DNA sample for comparison to the one that had been collected at one of the murder sites.
Holes ended up outside DeAngelo’s home after four months during which his team built family trees: On the basis of matches of second and third cousins, via GEDmatch, a team of five investigators, aided by genealogist Barbara Rae-Venter, put together a genealogical chart going back to the 1840s. On that day, the eve of his retirement, Holes saw DeAngelo only as a possibility, albeit an interesting one. In the end, however, he decided to leave without speaking to the suspect.
“I realized I didn’t know enough about him. If I had knocked on that door, things could have turned bad,” he told me in a conversation from his home in Colorado.
Holes’ experience exemplifies the fact that genetic genealogy is not a substitute for traditional police work. If he hadn’t gone on questioning witnesses, looking for DNA samples and seeking patterns in the killer’s mode of operation, the requisite evidentiary foundation for embarking on the final stage in the hunt for the killer would probably not have existed. “This is obviously a revolutionary technique,” Holes said. “But it won’t solve every case; nor should it be used in every case. It’s really almost like a last-resort tool.”
Anne Marie Schubert, the district attorney of Sacramento County in California, where DeAngelo, now 73, will face trial, also used the term “revolutionary” when I spoke to her about genetic genealogy.
Today Holes attends conferences of detectives and investigators, where he advises his former colleagues on the potential uses of genetic genealogy in their work.
The cofounder of GEDmatch, the internet tool that deserves credit for the recent wave of murder-case solutions, is Curtis Rogers, 80, a professional court-appointed legal guardian from Florida.
Until it emerged that his search engine had helped bring about the arrest of the Golden State killer, he had no idea that police were even using it. Rogers told me that reports of the suspected serial killer being arrested made him slightly suspicious. “I wonder if we were involved,” he said to his wife that day, April 25, which by chance also happened to be International DNA Day. The answer came a day later, when he arrived at his office and found TV-broadcasting vans waiting outside.
“It came as a big surprise, no question,” he said, although it was clear from his tone that the initial astonishment had passed and that he’s now proud of his role in solving the cases.
Rogers founded GEDmatch in 2011 together with John Olson, 67, an electrical engineer from Texas. Rogers still works almost nonstop, he says, but recently decided to retire from his day job and devote himself full-time to the website.
The basic idea of GEDmatch is simple. A commercial breakthrough in 2009 – the sale of autosomal DNA tests directly to the public – enabled genealogists to locate living relatives and not only those who died long ago, as earlier techniques had made possible.
“Suddenly you had this mass of people you knew you were related to,” Rogers says, adding that the enthusiasts began bombarding each other with emails in order to decipher the family ties between them. The process was long and tedious, and Rogers asked Olson to build an algorithm that would computerize the search. The two developed what turned out to be GEDmatch, and decided to invite the world to use it.
There are significant differences between GEDmatch and similar, commercial sites. First, Rogers and Olson offer their service for free, apart from a few features that cost $10 a month. The two men obviously operate the site out of love for genealogy. They also do not do DNA testing and don’t have a lab. Instead, users upload to their site the genetic data they have received from commercial testing firms. GEDmatch also makes it possible to upload DNA sample from every one of the firms, provided it comes in the right type of file, and to compare profiles, something that cannot be done in the case of people who were tested by different companies.
According to Rogers, visitors to the site were now apprised that the authorities had access to it and that they were invited to delete their information if they wished. “We became very open and honest, and we try to educate our people that that’s a possibility,” he says, referring to the possibility of law enforcement using the site.
Rogers was apprehensive about users’ reactions, following the arrest of the Golden State killer. After all, though the police were already using the site, those who had uploaded their DNA profile weren’t necessarily aware of that. So he was relieved to get many messages of support.
One of the more dramatic examples of such support came from Carol S., whose father murdered her grandfather in Detroit in 1979. Carol’s father had been in and out of mental institutions throughout his life. She told me that he had been tried for his father’s murder, found insane and sent to an institution again, this time for a lengthy period.
“I felt and I do feel that my father was responsible for many crimes,” she related. “I asked [the site]: How do I make sure that my GEDmatch profile is available publicly for law enforcement purposes.” She hopes her decision to allow the authorities to access her DNA profile will bring some comfort to victims’ families: “Suffering when your loved one is murdered is hard enough, and that’s why I feel it’s so important,” she says.
CeCe Moore wasn’t involved in the Golden State killer case, but the way it was solved made her uneasy, too. She now accepts Rogers’ new policy, and thinks that the major publicity the case received has changed users’ perception of privacy.
“Once everyone in GEDmatch was made aware that their data might be used that way, it really gave me the freedom to do my work,” she notes.
According to Moore, in light of the worldwide publicity the Golden State killer case has received, it can be assumed that everyone whose information appears on the GEDmatch site has voluntarily technically consented to allow the authorities to use it. Some experts dispute that and say that legal problems are liable to arise. But for CeCe Moore and GEDmatch, the decision has already been made.
“I had a lot of sleepless nights, when a district attorney would write to me asking for help in catching this serial killer, who might kill more people. I know I have the skills to do it, and it was very painful to have to say no” for ethical reasons, Moore relates.
Exploiting genetic info
Last October, the journal Science published a study done by a team headed by Prof. Yaniv Erlich, chief science officer of the Israeli company MyHeritage (also involved in the research were Dr. Shai Carmi, Prof. Itsik Pe’er and Tal Shor). The online genealogy platform, known here mostly thanks to the family trees that can be created through it, also has a genetic testing unit. In the wake of developments in America, Erlich and his team examined the genomic data of 1.28 million people and “project that about 60 percent of the searches for individuals of European descent will result in a third cousin or closer match,” according to the study. This figure is in keeping with the data reported by Parabon, and the MyHeritage team estimates that the number would rise considerably, as more people upload their profiles to GEDmatch. They also predict that with the aid of genealogical tools, it will be possible to locate almost every one of the search targets
In regard to Israel, Erlich told me that the MyHeritage genetic database includes only a few thousand Israelis, because regulations bar companies here from offering genetic testing for “recreational” purposes. The company, he says, will go to court to fight any order to open the database to the police, though that hasn’t yet happened. GEDmatch is of course accessible to the Israel Police, but is not especially useful as long as no Israelis are availing themselves of it. Asked about the services it may take advantage of in its work, the Israel Police told Haaretz that it doesn’t not answer questions about investigative procedures or cooperation.
Moore had reservations about the unequivocal character of the study that was reported in Science. In most cases, she says, it’s quite difficult to create a family tree and locate a target person. Adoption, migration, endogamy, common names, restrictions on document access or cases in which the biological parentage is ascribed to the wrong person – all these make it hard to follow the elusive DNA trail. Dealing with the April Tinsley case, with many variables, was “like fighting a battle through a dark jungle,” she recalls.
In any event, the study has stirred considerable interest in the U.S., where DNA databases are springing up rapidly. All together, the large commercial genealogy companies have about 17.5 million profiles. Haaretz was told by 23andMe that it has more than 5 million, Ancestry.com claims to hold 10 million, and MyHeritage reports 1.4 million DNA profiles. Family Tree DNA didn’t respond to an inquiry, but reports indicate that it has another million profiles. According to GEDmatch’s Roger Curtis, half of their million profiles are from the United States and the others from elsewhere around world, and the database is growing at the rate of 1,000 to 2,000 new users per day.
The question now is: What’s next? As things stand now, anyone with the means and the right skills can utilize GEDmatch to locate people of interest. Theoretically, both individuals and authorities in foreign countries will be able to do the same when the database contains profiles of enough of their citizens. The question of what will prevent U.S. authorities from exploiting the information in other cases – with respect to drug or property offenses, or, pessimists worry, even to track down holders of particular political views – remains unanswered. Is it a good thing that at police stations across the United States they are gradually building maps of people’s family relations, including relatives of whose existence those people are unaware?
Prof. Erlich, from MyHeritage, noted that the motivation behind the study his team conducted was the desire to draw the world’s attention to the developments in the realm of genetics and to offer solutions for supervision and legislation that will protect users’ privacy. We are only at the start of a revolution in genetics, says Erlich, who predicts that entire population groups will undergo genetic sequencing. “What’s happening now is the building of the foundations for the future, which will serve us for the coming decades.”
Cynics might detect an economic threat to the commercial companies if the police suddenly depart from the limits of the public consent they received because of the Golden State killer case, and the advantage MyHeritage may have in being seen as a leader in the struggle for privacy. Still, it’s impossible to ignore the fact that information databases leak, that DNA is a powerful tool and that Erlich’s comments carry weight.
“There are huge companies in Silicon Valley that say the writing was on the wall for years and nothing was done,” he notes. “So, if the writing is on the wall, let’s be practical and try to see how we can help this field and move it forward. What we’re trying to say is that it’s not too late yet.”
One idea proposed by Erlich and his colleagues is to develop a cryptographic key that will mark every DNA test carried out by one of the large commercial companies, and to have GEDmatch accept only profiles with that marking. Anyone who tries to upload an unmarked profile will have to answer questions relating to his identity. Another idea is to impose restrictions on law enforcement authorities when they try to collect genetic information.
Echoes of Erlich’s forecasts are also discernible in the views of Erin Murphy, a New York University law professor who specializes in technology and forensic evidence in the criminal justice system. “We’re hurtling into a world where genetic information is going to become incredibly useful and powerful for employers in regard to life insurance, health insurance and so forth,” she told me. “The bottom line is that it will be cross-applicable because it’s the genome.”
More concretely, there is a concern that at some point, various bodies – whether governmental or private – will make decisions about our lives on the basis of genetic information we provided unthinkingly.
Anne Marie Schubert, the Sacramento district attorney, says she has built a model for supervising the use of genetic genealogy, similar to the one used to oversee the FBI’s database. (To gain access to the latter, police officers must sign a statement declaring that the crime under investigation was violent and that all other possible avenues of investigation have been exhausted.) Murphy sees this as a step in the right direction, but adds that far more comprehensive legislation is required.
“I think people fail to appreciate exactly what law enforcement is doing when they use those searches,” adds Murphy, “what’s called SNP testing [used by the commercial testers for recreational genealogy]. These are really the keys to a person’s identity and existence.”
On this point, Paul Holes, the retired detective, notes that the police can only have access to matches between profiles, and do not get access to the complete genetic code itself.
“Why can someone I don’t even know exists give away my genetic information?” she asks, adding that there are legal precedents supporting this stance. “If you’re a joint occupant of a space – for example, I live in my home with my husband, and police show up to search the place. I tell them to go ahead and my husband says no. My go-ahead can’t override my husband’s no. I think we can think of the genome in the same way.”
Finally, Murphy explains that there is very little supervision in the field of genetics in the United States, and says that she supports the idea of a comprehensive regulatory structure in regard to genetic information. Given that political battles will probably prevent agreement on such legislation, she says that “the likeliest outcome is that this Wild West situation will continue. We will also see more of those cases [of crimes being solved by means of genetic genealogy], because one of the major impediments now is cost and technological limitation,” barriers that are constantly dropping. And because it’s unlikely that people will drop their leisure-time interest in genetics, she adds, they must at least think more seriously about how the law enforcement system uses genetic information.
Moore has a different take. “Most family members said you can never find closure. But they want justice. They want answers – and I can give them that,” she asserts, and goes on: “Maybe some of the people who commit these crimes will realize that it will be a lot harder to get away with them in the future. Maybe it will save one life, maybe it will save hundreds of lives. It’s worth it.”