In 1987 someone brutally murdered 20-year-old Jay Cook and his 18-year-old girlfriend, Tanya Van Cuylenborg. The case went cold until this week, when police in Seattle arrested 55-year-old William Earl Talbott II. They reportedly found Talbott using the same methods that led to the capture of the alleged Golden State Killer, Joseph DeAngelo.
The double homicide became famous due to Unsolved Mysteries, which aired a chilling segment about it in October 1989. Cook and Van Cuylenborg traveled from Victoria, Canada on November 18, 1987 to Seattle. They disappeared after boarding a ferry, only to be found miles apart later that month. Tanya was raped, and both were beaten and strangled.
Regarding Talbott’s arrest, KOMO quoted Snohomish County Sheriff Ty Trenary, who said crime scene DNA “was used to identify his ancestors which in turn led us to the identification of Talbott.” Just as in the Golden State Killer investigation, investigators used GEDMatch, a self-serve genealogy database to which users can upload raw genetic data in hopes of finding distant relatives.
Cold case detective Jim Scharf from the Snohomish County Sheriff’s Dept. reportedly said that “genetic genealogy that was the key tool that got this case resolved.”
“Had law enforcement never had access to genetic genealogy,” said Scharf, “I don’t believe this case would ever be solved.”
After they identified him as a suspect, police reportedly were able to obtain Talbott’s DNA from a cup he’d left in a work vehicle.
April Tinsley vanished on Good Friday, 1988. According to the FBI page on her disappearance and murder, the 8-year-old Fort Wayne, Indiana girl was on her way home from visiting a friend when she was kidnapped. April’s killer raped then murdered her. She was suffocated to death.
April’s killer has never been caught. Police have a description, an approximate age range, and a psychological profile indicating he is likely a “preferential” pedophile — a pedophile who is specifically attracted to children, likely within April’s age range. There are different kinds of pedophiles, according to most profiling research, and this kind can find it harder than others to hide their attraction to children.
The man who killed April Tinsley even left behind plenty of evidence. Including communications, beginning with a message scrawled on a barn door, reading in part, “I kill April Tisley (sic)… I kill again.”
Michelle McNamara wrote about April’s “communicative” (Michelle’s apt description) killer in a blog post at True Crime Diary in 2012:
Investigators say they believe whoever wrote the message was April’s killer. They haven’t said why, but there are two possibilities. The writing implement, said to be crayons, was found nearby, and DNA from the crayons could have been matloched to a sample found on April’s body. The second possibility has been discussed on message boards, but not confirmed by investigators. It’s alleged that when April’s body was found she was fully clothed but missing one shoe. The rumor is that above “ha ha” was the question, “did you find the other shoe?”
As Michelle reported in that post, 7-year-old Sarah Bowker would disappear not long after that message on a barn door. She met the same fate as April. Strangely, while the coroner who examined both girls concluded their murders were related, the FBI disagreed.
The Tinsley murder had long been a cold case by the time April’s killer made himself known again in 2004. He did this by leaving seemingly barely literate notes inside baggies, accompanied by used condoms in various areas, clearly targeting little girls anew.
Police believe he is a white male currently in his 40s or 50s who prefers and desires sexual contact with children, particularly little girls.
“This offender has demonstrated that he has strong ties to northeast Fort Wayne and Allen County,” the profile said. “This is where he likely lives, works and/or shops. You may be standing next to him in line at the grocery store, sitting beside him in the pew at church, or working beside him on the production line.”
Such profiles can be helpful in that they might spur local residents to tell police, “You know, I always wondered about this one guy (…)”
Criminal behavioral analysis — profiling — is a fascinating if imperfect art. In development since the 1970s, it’s accrued an aura of myth, thanks in part to great fiction like Silence of the Lambs. The truth about behavioral analysis is that it’s one tool in a vast investigative suite of them, and is rarely the factor that solves the case — it’s simply one helpful way to narrow down a field of suspects.
Parabon is an outfit in Virginia and they’ve been doing a steady business creating suspect portraits using an unknown subject’s genetic history as revealed through DNA. Here’s how they describe what they do with a product they call Snapshot:
Snapshot is a revolutionary new forensic DNA analysis service that accurately predicts the physical appearance and ancestry of an unknown person from DNA. It can also determine kinship between DNA samples out to six degrees of relatedness. Snapshot is ideal for generating investigative leads, narrowing suspect lists, and identifying unknown remains.
Parabon, as best as I can tell, uses the same data that Ancestry.com and 23andMe utilize from submitted DNA samples to develop a portrait of unknown killers as well as John and Jane Does. It’s a smart and understandable use of new technology and when I first learned of it, I was blown away. Anyone who has ever obsessed over a cold case in which there were known samples of the killer’s genetic material hears about what Parabon is doing and gets a little thrill at the prospect: What would a Parabon Snapshot of the Zodiac Killer look like? Or the vicious Golden State Killer, whom Michelle McNamara was writing a book about when she passed away?
My wife and I have had our DNA analyzed by both Ancestry and 23andMe. It wasn’t until I took a deep dive into what 23andMe concluded from my DNA sample that I thought about Parabon’s Snapshot profiles and felt a sudden twinge of disappointment.
As Parabon attempts to make clear in profiles such as the one they’ve released in the Tinsley case, their workups — highly detailed suspect sketches, basically — are based on the suspect’s most probable appearance, based on what their genetic material tends to predict.
Here’s the cold water about this kind of thing: even if your genetic material tends to indicate you’ll look a certain way, there is no guarantee you will. I learned this from my 23andMe experience.
23andMe breaks reports from your DNA down into numerous reports. These include:
Ancestry Composition (I am 99.2% northern European and 0.7% Sub-Saharan African)
Muscle Composition — I am a “likely sprinter,” my muscle composition primarily “fast-twitch” muscle fibers
Caffeine Consumption — I’m likely to consume less. I do not consume less. I consume large quantities of it.
Individual reports on your likely hair color, physical characteristics, and skin.
The reports in the last bullet point are where things get interesting.
According to 23andMe’s analysis of my DNA, I am most likely to have “light brown or blond hair.”
On my hair report, 23andMe actually states, “You are not likely to have red hair. 94% of customers who are genetically similar to you do not have red hair.”
I was born with coppery red hair. It’s gotten blonder as I’ve aged, but there’s no doubt about it. The report nails my hair type (lightly wavy), that I’m likely to be balding (I am), and the fact it’s light colored. But that’s it.
It seems like nitpicking but if the same data was used to make an unknown suspect profile, he wouldn’t have red hair — a highly distinctive feature.
My 23andMe profile correctly predicts I have light-colored eyes (green), but the part of the report that details potential facial characteristics of someone with my genetic makeup states clearly: “Steven, you are not likely to have a cleft chin.62% of customers who are genetically similar to you do not have a cleft chin.”
One of my most obvious facial characteristics is a clearly cleft chin.
23andMe also concluded I would have light skin, which I do, but not many freckles. That part is debatable. I don’t have nearly as many as some redheads, but I’ve certainly got some.
So far, it’s not hard to guess that a Parabon Snapshot reaching conclusions similar to 23andMe from my genetic data would have me — a redhead with some freckles and a cleft chin — as a light brown-haired man with few freckles and a square, un-cleft chin. My DNA tends to predict those things. I simply was an outlier and ended up born with the less likely traits. Just as many people are, every day, everywhere.
This is not a knock on Parabon. I think they’re doing very necessary work and hope they keep refining and improving their product. It’s a reality check for the many true crime devotees like me who see intriguing news stories about Snapshot-generated profiles and don’t bother to look past quickly turned-out reports breathlessly hinting that this might the thing that solves the case.
It’s a long-needed and possibly fundamental tool. Truth be told, I have no doubt a Parabon Snapshot or some similar forensic product will one day play a key role in solving a major case, especially when combined with a service that has Ancestry.com’s capacity for identifying possible familial relations, up to 3rd, 4th, 5th cousins.
But Parabon’s Snapshot is not a magic bullet. Reading the fine print on Parabon reports it is clear they are aware of this. Reading interpretations of their work in the press, it’s clear that the media is not. Sober, realistic assessments of possible breaks in long-unsolved and deeply unsettling cases like the murder of April Tinsley don’t traffic quite as well for a TV station or newspaper website.