How DNA technology helped police identify the alleged Fairmount Park rapist

Philadelphia police spent years slowly closing in on a suspect in the Fairmount Park rape cases — employing DNA technology that’s increasingly used to solve cold cases — before he was arrested this week for an unrelated crime.

Philadelphia Police identified 46-year-old Elias Diaz on Tuesday as the so-called Fairmount Park rapist. He is accused of assaulting four women, killing one, between 2003 and 2007. Diaz was arrested over the weekend for allegedly riding through Pennypack Park with a machete attached to his bike and slashing people. He has been charged with murder in the Fairmount Park death, and investigations into the other Fairmount Park assault cases are ongoing.

Investigators around the country are turning to forensic genealogy for leads in cold cases, using the public’s affinity for commercial genetic testing to hunt for the relatives of unidentified killers and victims. The technology was used to identify the Golden State Killer in California in 2018, and to investigate at least two other high-profile cases in the Philadelphia region in the last year.

Philadelphia Police worked with Parabon Nanolabs, a leader in the crime-solving science known as forensic genealogy, to comb commercial genetic databases for new clues in the Fairmount Park assaults.

“Detectives come to us when they have an unsolved case where there is DNA, but the profile has not matched anyone,” said Ellen Greytak, Parabon’s director of bioinformatics. “They’re looking for ways to get other leads from the DNA.”

DNA evidence collected from the crime scenes initially produced no matches in national crime databases, police said. With Parabon’s help, authorities identified Diaz as a person of interest earlier this spring. Still, no one knew where he was.

Even when it seemed likely that Diaz might be dead, investigators kept working to find him, said Ryan Gallagher, the assistant director of the Philadelphia Police Department’s office of forensic science.

“There were still victims out there that would want to have closure,” he said. “Everyone was shocked when we found out that he was still around.”

Here’s a look into how law enforcement agencies use forensic genealogy and other new DNA technology.

Parabon helped police zero in on a suspect in the Fairmount Park assaults by performing a genealogy search to try to find family members who could lead police to the attacker.

Forensic genealogists compared a DNA profile of the suspect against public databases where millions of people have uploaded their own DNA profiles, gleaned from commercial genealogy tests like 23andMe or Ancestry.com. (People have to opt in to public databases and allow law enforcement to access their information.)

The search produced a list of everyone who shared DNA with the suspect. Investigators used it to reconstruct the suspect’s family tree.

Most people share 17% to 34% of their DNA with each grandparent, Paul Woodbury, a genealogist at Legacy Tree Genealogists in Salt Lake City, said in an interview with The Inquirer last year. Half-siblings share about 25% of their DNA with each other, and first cousins share 4% to 23% — meaning that a 20% match in a genealogical database could indicate a suspect’s first cousin, half cousin, grandparent, or grandchild. That’s when investigators start combing through birth and death records and interviewing family members.

Investigators in the Fairmount Park rapist case had found fairly close relatives to the suspect through the DNA database search. But the suspect had a large family; police had to work through a family tree of almost 1,000 people with “no clear associations to the city of Philadelphia,” Gallagher said at a news conference Thursday.

In some cases, police worked with the FBI and approached family members to ask them to submit DNA to help narrow the search.

“We had the door slammed in our face a couple times, figuratively,” Gallagher said. “Other people were very helpful.”

Eventually, the investigation concentrated on one branch of the family.

“We had the family name, and there were multiple people in that family that could have been a potential suspect for us,” Gallagher said. “But there was no clear picture there until Elias popped up.”

Investigators had identified Diaz by name before he was arrested, but family members who had been estranged from him for years were unclear about his whereabouts, Deputy Police Commissioner Frank Vanore said Tuesday.

Parabon also created a “DNA phenotype” that helped develop a composite sketch of the suspect based on his genes. The company used DNA to predict the man’s ancestry, eye color, hair color, skin color, freckling, and face shape, Greytak said.

In the past, an artist may have produced a composite sketch based solely on descriptions from witnesses, who told police they had seen a Hispanic man with black hair, a widow’s peak, and prominent eyebrows after the Fairmount Park rapes. Parabon’s DNA profile added more detail to the rendering.

In 2021, Parabon sent three sketches to Philadelphia Police that were then released to the public.

Genetic genealogy has led to breaks in several cases in the Philadelphia region in the last year. Last December, a third and fourth cousin who uploaded DNA information to a public database helped identify Joseph Augustus Zarelli as the murder victim known for decades as the “Boy in the Box.”

The technology also led to an arrest in July 2022 in the long-unsolved murder of Lindy Sue Biechler in Lancaster County. In that case, genealogists from Parabon found that Biechler’s killer’s DNA matched that of people who had ancestors in a small town in Italy, working through generations of records to find a suspect.

Still, using DNA and genetic databases to profile suspects has drawn criticism from some scientists and privacy advocates.

A 2020 article in Nature noted that scientists are still unsure how a person’s genetic predispositions to, say, blue eyes or brown hair combine to “become a unique human face.” Others have raised privacy concerns about using commercial DNA databases, which are largely unregulated, WIRED reported in September.

Staff writers Tom Avril, Chris Palmer, Ellie Rushing, and Rodrigo Torrejón contributed to this article.