A Little Girl Went Missing In 1988. Law Enforcement Was Stuck For Years. Then Parabon Got Involved.

But this work has also set the company down a path with its own unknowns, riddled with privacy concerns, competitive threats and ethical questions, as Parabon methodically chips away at hundreds of thousands of cases — like that of little April Tinsley — with the help of largely unregulated genetic research.

“It can be extraordinarily powerful,” says Ellen McRae Greytak, Parabon’s director of bioinformatics. “But it took a long time to convince people of that.”

Among the many questions that swirled around April Tinsley’s murder, her autopsy made one horrifying determination pretty fast: The first-grader was sexually assaulted and suffocated to death.

Two years later, other neighborhood girls found a disturbing crayon-scribbled confession of her murder on a barn wall. And 14 years after that, the killer planted menacing notes inside bags, alongside used condoms, atop bicycles and inside mailboxes. 

With those threats, he left behind his DNA.

That was nearly a decade before Parabon first opened its doors as a software firm, helping pioneer cloud computing 550 miles away in Virginia. And it was a full 15 years before the company upended its business model.

That began in 2005, when CEO Steven Armentrout spoke at a Marshall University conference about his company’s computational tools. There, he met Mike Norton, a Marshall professor who told him about an emerging field called DNA nanotechnology.

“I had no idea in the moment what we could build with it. But I knew if you could manipulate matter at the nanoscale, and make trillions of identical copies of molecules in a programmatic fashion, that the potential was huge,” Armentrout says. “And we’ve been pursuing that ever since.”

The DNA nanotechnology software builds smart drugs, tiny structures also known as nanosensors that carry therapeutics and detect specific molecules in diseases like cancer. It also led the new partners to create the Parabon NanoLabs division, which quickly outgrew its parent and now functions as the core business, encompassing both therapeutics and forensics. Parabon still does cloud computing, Armentrout says, but “only to the degree that it supports our own work.”

That work moved deeper into DNA technology in 2010, when the company won a Small Business Innovation Research program solicitation from the Defense Department’s Defense Threat Reduction Agency, which was seeking phenotype prediction capabilities — to ascertain physical appearance from genetic information — to identify people responsible for planting improvised explosive devices in the Middle East.

That led to a 2012 phase 2 project for Parabon, which brought on McRae Greytak, armed with a Harvard University doctorate degree in evolutionary genetics, to lead its bioinformatics team. The charge? Figure out if it was possible to predict what people look like from their DNA. She built the tool — Snapshot — to construct a composite sketch of a person with attributes ranging from face shape to pigmentation and individual characteristics down to eye and hair color. 

Then came the idea to start using that technology to help law enforcement track down criminals.

“It was a slow burn at first,” McRae Greytak says. “We would go to conferences and we would tell people, ‘We can predict what your suspect looks like,’ and they’d say, ‘No you can’t.’” 

---

Parabon began using a DNA marker called a single nucleotide polymorphism, or SNP, which carries more genetic information than the short tandem repeat, or STR, used in traditional forensics — including the FBI’s Combined DNA Index System, or CODIS, database. Simply put: The SNP was much more comprehensive.

But Snapshot had its critics. Worse yet, people didn’t buy it. Labs wouldn’t test Parabon’s small samples, skeptical it could be done. Law enforcement agencies put Parabon through the ringer, requiring auditions with fake cases before they’d agree to work together for real. Forensics experts dismissed it, doubting its potential entirely.

“In retrospect, had we known all of the things that we would have to overcome, we wouldn’t have started — it was just that daunting,” Steven Armentrout says. 

“It was just frustrating that it took so much energy and so much convincing,” he says. In theory, “we understood the instruments could operate on these scales — we had some things to prove to ourselves. And it wasn’t a matter of whether, it was the degree to which we were going to be successful. But we knew we could provide information that they didn’t have. And it turns out that is the case.”

In mid-2014 — the year April Tinsley would have turned 34 — the investigation took a sudden turn.

The Fort Wayne Police Department got wind of Parabon. The department shared the killer’s DNA left in one of the baggies, and the company applied its Snapshot technology. It returned a composite sketch of a suspect to the police, the first time it had done so for law enforcement.

A new door opened in an otherwise impenetrable concrete wall.

Things opened up for Parabon, too. It picked up more police cases. It also landed another DOD contract in 2016 to improve the algorithms that the agency’s Armed Forces DNA Identification Laboratory used to connect service members’ skeletal remains to their family members. To do that, Parabon needed to recruit large families willing to share their DNA, so Vice President Paula Armentrout spent weekends sitting at a card table in area neighborhoods, looking for genealogists — often hobbyists who trace their family trees. 

Folks she encountered suggested she speak with genetic genealogists, those specifically using DNA to find unknown relatives. Those conversations led her to CeCe Moore, a prominent genetic genealogist known for her work on unknown parentage cases and the PBS series, “Finding Your Roots.” Moore had helped hundreds of adoptees find their birth parents, and her DNA Detectives Facebook page alone counts more than 120,000 members. 

Parabon convinced her to join its staff as head of its new genetic genealogy unit in May 2018. Despite her few law enforcement connections, Moore was the missing piece to Parabon’s genetic genealogy puzzle.

And it snapped into place about a month after another pivotal, parallel maneuver: When authorities on the West Coast caught one of the most prolific serial killers in history. 

As Indiana police still chased leads for April Tinsley’s murderer, California authorities identified the Golden State Killer, who committed more than a dozen murders, 50 rapes and 100 burglaries from 1974 to 1986. On April 24, 2018, investigators identified him as Joseph James DeAngelo, a former police officer and Navy veteran, thanks in part to a genetic genealogist who gleaned information from the killer’s DNA and a public genetic genealogy database — the first time such a database was ever used for that purpose.

That database, called GEDmatch, was created by Curtis Rogers, now 81. He launched it in 2010 with friend John Olson for fellow genetic genealogy hobbyists who could now find unknown relatives by scouring newly available autosomal DNA — which reflects generations worth of people who have passed along DNA, not just a single female or male line. Rogers and Olson decided to make it completely available to the public.

That evening in April, Rogers and his wife were listening to the news with a glass of wine when a story popped up about a cold case heating up because of advanced genealogy. The next morning, he received an email — from CeCe Moore, though unrelated to Parabon at the time — that GEDmatch was involved. She had recognized the technique from her earlier work.

“That was a shock,” Rogers says. “By the time I get to my office, … there are all these satellite trucks out front, and there were reporters walking around wanting to get a statement from me.”

It was a pivotal moment for Parabon — and proof that it should go live with its own genetic genealogy service; if the technology could remove one of the most atrocious criminals from the street, then it was worth pursuing. Within 72 hours, Parabon had made its next business decision.

“We all looked at each other and said, ‘We’re going to do this. We’re going to take the necessary steps to make this offering available,’” Steven Armentrout says. “We were sitting on these digital files that we could’ve, had we chosen, uploaded to GEDmatch at any time — of course, with the police’s permission. And it was a matter of committing to doing that.”

So Parabon did, but at considerable internal cost to its legacy phenotyping product.

“We cannibalized our own business,” Armentrout says. “We turned our business completely on its head and changed all of our collateral, marketing, came up with cost structures — all of it, we did in two weeks’ time.”

In May 2018, Parabon circled back to the Tinsley investigation. This time, the company could apply genetic genealogy to that DNA evidence.

Moore spent about nine days working the case, breaking down a complex web of links to the crime scene DNA. She set the bar low, expecting to come back with a common ancestor like a cousin or grandparent, perhaps a family name. But she ended up whittling the genealogy search down to two brothers — who were still alive. The Parabon team set up a briefing with police, who had spent years chasing different leads.

“We were floored. All these years, 30 years in the roller coaster that we’d been on, up and down, and thinking we had the guy, ... only to have our legs kicked out from under us,” says Detective Brian Martin of the Fort Wayne Police Department. “We just looked at each other, dumbfounded.”

But Parabon just had names. Police still needed to conduct their own investigation, and to confirm a direct match between suspect and crime scene DNA.

The genealogy helps police find a suspect, “but it is not what’s going to be used to convict that person,” McRae Greytak says. “In the vast majority of cases that we’ve worked on where it’s gone to court, the person has plead guilty because there is that DNA match.”

It can’t lead to a wrongful arrest, either, Moore says. “It wouldn’t happen. If I pointed toward the wrong person, they would never get their direct match.”

“We’re not putting people in jail because Parabon NanoLabs says, ‘John Doe did it.’ Parabon NanoLabs says the same thing as the store manager who says, ‘I saw so-and-so do it.’ Well that’s great, but now we have to corroborate that and we have to be able to prove beyond a reasonable doubt that that, in fact, did happen,” says Sgt. Jacklyn Davis, spokeswoman for the Anne Arundel County Police Department, which has also worked with Parabon, including on a 60-year-old homicide cold case.

Parabon’s work has been featured in The New York Times and The Wall Street Journal, increasingly catching public interest because of growing popularity around the true crime genre and forensic work — thrust into the spotlight by binge-worthy TV docuseries like “Making a Murderer” and podcasts like “The Murder Squad.” 

But concerns over the technology and practice still swirl around the company. In the wake of the Golden State Killer breakthrough, some genetic genealogy databases shut down because their creators didn’t want to be a part of it. A steady stream of news stories questioned Parabon’s methods and accused it of invading people’s privacy by quietly using recreational genomics websites like Ancestry.com and 23andMe; Parabon says it does not. Crisis media management quickly became part of the Parabon team’s day-to-day work, as articles claimed Parabon accessed people’s private genetic data — claims Parabon refutes. It says it can’t access those consumer sites, let alone specific genetic information, if it tried.

“All you can see is how much DNA they share with the sample that you uploaded,” McRae Greytak says of databases like GEDmatch. “You can’t go look at DNA and then say, ‘Do they have this mutation that’s associated with this disease?’ You cannot do that.”

Add to that the complexities of genealogy. Despite never taking a DNA test, a person could be linked to a distant relative who uploaded and made public his or her own DNA test data to a GEDmatch or FamilyTreeDNA — which now, too, enables these searches. That’s led some to question the legality of an industry without a legal precedent. Some states have introduced legislation to regulate such practices while others are pushing bills to formalize it. Steven Armentrout has found himself testifying before the Maryland General Assembly’s House of Delegates judiciary committee, which was considering legislation that would ban police from using such databases to trace crime scene DNA, after the state outlawed what’s called familial search — traditional DNA matching to find parent-child relationships in criminal databases — more than a decade earlier.

Armentrout says the two tap two different sources — traditional, government-owned criminal databases for familial search and publicly available databases with volunteered samples for genetic genealogy. “My entire testimony was trying to point out how wildly different these two things are.”

House Bill 30, legislation sponsored by Del. Charles Sydnor, D-Baltimore County, didn’t advance earlier this year, a relief for companies like Parabon that defend the value of the technique. But, Armentrout says, “it’s going to get resurrected again, I gather.”

Genetic genealogy has its own limitations. Parabon can enter DNA of an unidentified person into GEDmatch, but if none of that person’s family has done DNA testing or uploaded it to the database, “you’re not going to be able to work on that case,” McRae Greytak says. “And you don’t know that ahead of time until you’ve gone through the work of doing the lab work, putting it in the database and checking.”

Ultimately, in what’s a very decentralized criminal justice system, “the technology is way ahead of the implementation,” says Gary LaFree, chair and professor of University of Maryland’s Department of Criminology and Criminal Justice. “You want that magical balance between catching people doing awful things and having some modicum of privacy.” 

The U.S. Department of Justice is taking that on. It issued an interim policy on forensic genetic genealogy this September to establish guidance for law enforcement in using the tool for investigations, including the criteria required to use it. The goal, according to the DOJ, is to balance public safety and privacy interests. It’s slated to go into effect Nov. 1, with a final policy to be issued in 2020.

“It seems like a fine set of guidelines,” Armentrout says. “It doesn’t seem to be particularly onerous to us.”

The investigation took 13 days after Parabon supplied police with the names of two brothers as potential suspects in April Tinsley’s case. On July 15, 2018, police arrested one of them, John D. Miller.

Detective Martin and his team began 24-hour surveillance on Miller after their own research pinpointed him as the top suspect. They visited and covertly obtained garbage from his home — used condoms, a soda can and TV dinners — and sent the DNA from that trash to a crime lab. It matched that of the used condom he’d left in the bags for other girls years prior. And it matched the DNA left on Tinsley’s body and clothing. 

Martin and his partner took Miller into custody and, with the evidence against him, got a confession. Police confirmed with another test that his DNA matched that of Tinsley’s killer.

It was him. They got him.

Miller, then 59, ultimately pleaded guilty to charges of murder and child molestation. His conviction was quick, but his sentence is for 80 years in prison.

“It’s very humbling to be a part of something like this, to be able to give a family answers,” says Martin, a Fort Wayne teenager himself at the time Tinsley was killed.

“You’ve got to think all the way back to ‘88,” he says. “The crime scene technicians and detectives that did their jobs back then enabled us to be able to go forward 30 years later, and here’s new technology to bring this home. That is just an awesome thing to be part of.”

In the year since that confession, Parabon has grown dramatically as well. It’s more than doubled its employee count to 26. Its leaders wouldn’t disclose its revenue, but said they’re finding new routes to earn it, applying its technology to more arenas. The National Institutes of Health recently funded its effort to develop a genetic test to detect the risk of Alzheimer’s disease. Parabon’s nanotechnology side has two NIH grants to advance an HIV vaccine and, in other areas, use DNA as a substrate to build molecular devices and therapeutics. 

And there’s the flip side of the police work. Moore says Parabon’s genetic genealogy team could use its approach to exonerate innocent people. “I would very much like to help free someone who’s behind bars wrongly,” she says. “I want to keep doing what we’re doing, but I’d also like to work some of those cases where that’s actually our purpose — where we’re perhaps hired by the defense, a post-conviction trial attorney.”

In the meantime, Parabon is hoping to keep closing cold cases as it did with young April Tinsley’s murder — a case that took 30 years, three months and two weeks from that devastating day she disappeared.

---

New federal guidelines

The Department of Justice released interim guidelines Sept. 24 on how law enforcement should use forensic genetic genealogy (FGG) for criminal investigations — in hopes of balancing its effectiveness with the need for privacy. Ultimately, departments need to exhaust their other methods first and document wherever they use FGG technology. Here’s more information:

The users:

• Agencies within the DOJ or those that receive DOJ funding for conducting criminal investigations
• Any governments that receive grant dollars from the DOJ to conduct forensic genetic genealogy
• Department employees conducting research on leads from the technique

The limitations:

• Information from FGG is to be used only as an investigative lead; an agency can’t arrest a suspect based solely on a genetic association from this service.
• To use that lead, a direct match must be made between the individual’s DNA and a crime scene sample. The suspect’s DNA profile must be directly compared to the profile in the Combined DNA Index System, or CODIS, an FBI database with DNA indices for known offenders. “This comparison is necessary to confirm that the forensic sample could have originated from the suspect,” the guidelines read.

The criteria:

• An agency can consider FGG in an unsolved violent crime with a forensic sample from a perpetrator or to help identify remains of a suspected homicide victim.
• An attorney may also allow its use for other violent crimes or attempted crimes beyond homicides or sexual offenses when there’s a threat to national security or public safety.
• An agency must first upload the sample to CODIS and fail to produce a DNA match.
• An agency must contact a CODIS lab official to determine the sample’s viability for FGG.
• The technique shouldn’t be used by law enforcement agencies, labs, genetic genealogy companies or others to determine a person’s predisposition for disease, medical conditions or psychological traits.

---

The competition

Lorton player joins the emerging field

Parabon may have been a pioneer in its field, but other companies were fast to follow.

One of them is in Parabon’s back yard, Lorton-based Bode Technology. The 25-year-old forensic DNA testing company works with crime labs, prosecutor’s offices and law enforcement to sift through backlogs of sexual assault kits and unidentified remains work, predominantly supporting investigations through the FBI’s CODIS database. It launched its forensic genealogy service earlier this year.

“We know that as these profiles get into the CODIS database, about half of them result in investigative leads or matches within the database, but half of them do not,” says Andrew Singer, vice president of global marketing and operations sales for Bode, who oversaw the new service’s rollout. “That’s where the role of forensic genealogy came into play.”

Bode’s total operation is much larger than that of Parabon. About 200 Bode employees, including 140 scientists, work in its DNA testing line, which makes up approximately 70% of the company’s total revenue, according to Singer, who declined to disclose specifics.

And Bode is no stranger to Parabon. Before starting the new service, Bode provided some of the initial DNA sample extractions for Parabon’s cases, and the two have collaborated on research, funded by the National Institute of Justice, on how to best use the technology for unowned DNA samples.

Others, too, increasingly enter the game, from private companies to independent genealogists to individual law enforcement agencies. 

But, Singer acknowledges, “Parabon is our biggest competitor.”

Original Article