Colorado D.A. offers Philadelphia help in Kensington strangler case

Denver District Attorney Mitch Morrissey stands next to a model of DNA structure in his office in Denver. (File Photo / AP)

Denver District Attorney Mitchell Morrissey says he wants to help Philadelphia police catch the Kensington strangler using a new forensics technology known as familial DNA.

The technique can help zero in on suspects by using DNA to identify their close relatives.

Familial DNA has helped police crack several cold cases in England; last summer, it helped capture a suspected notorious California serial killer.

So far, the Kensington strangler has killed three women. Samples from the crime scenes show a common DNA signature, which could be used to seek relatives.

"When I read that they connected a third woman's death to this same DNA profile, I couldn't just sit on my hands," Morrissey said.

The technology needs special software, and Morrissey has offered Colorado's version for free, even offering to send experts here to install it.

He said he made the offer in e-mails to Philadelphia District Attorney Seth Williams, as well as to Police Commissioner Charles H. Ramsey, Fraternal Order of Police president John J. McNesby, and Mayor Nutter.

Williams and McNesby responded, Morrissey said, promising to forward the offer to detectives.

"I'm aware of the technology, and I've personally reached out to the FBI," Deputy Police Commissioner William Blackburn said. "We're going to explore the possibility of using this."

Scientists who have studied familial DNA say Philadelphia should take Colorado's offer seriously.

"This is exactly the kind of case that they should consider the familial searching method for . . . assuming they have a good, complete DNA result, and they've exhausted other reasonable leads, and there's an ongoing threat to public safety," said Frederick Bieber, a geneticist at Harvard Medical School.

The technique can sift through databases of known offenders and find people likely to be close relatives of a person whose DNA is found at a crime scene.

Some critics see potential civil-liberties violations. The technology can identify people with family ties to a sample at a crime scene, said Sheldon Krimsky, a Tufts University professor who studies science and ethics. "All of a sudden, these people's families are suspects," he said, even if they do not have any other ties to the crime.

Former Assistant U.S. Attorney Chris Asplen disagreed, saying that family members who come up in DNA searches are not necessarily suspects but could help lead to one.

Familial searching is not that different from other investigative methods, he said. If a witness who sees a getaway car identifies a partial license plate number, the police might search a database and get several hundred leads. "That wouldn't make everyone who came up in that search a suspect," Asplen said.

"The police would be searching data from known offenders," said Asplen, former DNA unit director of the National District Attorneys Association and a former deputy district attorney in Bucks County. "They're in the database for a reason."

DNA has been used for years to search for likely family relationships - starting in the 1980s, when geneticist Mary-Claire King devised methods to help identify children stolen during Argentina's "dirty war" and reunite them with their biological families.

One complicating factor is that familial DNA is not as precise as the standard DNA fingerprinting. That technology is extremely reliable.

The odds of a false match are minute because scientists use only segments of the DNA that tend to vary from person to person, and then compare at least 26 such regions, 13 from each parent.

With familial DNA, there can be false positives - people whose genetics suggest they could be close relatives when they are not.

But there are ways to sort through that, said Harvard's Bieber. Parent-child relationships are clearer-cut, always matching in 13 of the 26 sites tested, because children inherit half their DNA from each parent.

Siblings are more variable, Bieber said, as the contributions from each parent can be shuffled in different ways.

Bieber and two colleagues wrote a computer program that showed how to pick out the most likely direct relatives - siblings, parents, or children. They tested the program on known relatives and showed that it can work.

In California, the crime lab workers devised their own software, identifying the top 150 most likely relatives for a given sample.

They used it in their search for a Los Angeles-area serial killer of at least 10 people known as the "Grim Sleeper," so named for the long hiatus he took from killing.

Just as in the Kensington strangler case, the police in California had a good sample of the killer's DNA.

Their strategy was to take the top 150 likely family members of the perpetrator and further screen them with markers on the Y chromosome, which determines maleness and should be identical in brothers, fathers, and sons.

The first time California police tried this, they ruled out all 150 people the software identified. A year later, they tried again, and the Y chromosome markers matched a new sample from a young man recently arrested on gun-related charges.

That man had a father and uncle living in the area where the murders were committed.

Focusing on the man's father, police found his favorite pizzeria and dressed as waiters. After the man came in to eat pizza with his girlfriend, they bused his table, put his plate in an evidence bag, and had the crime lab test the discarded pizza crust for DNA.

They found an exact match with the Grim Sleeper.

Of course, that will work in the Kensington case only if the strangler has a close relative who has been convicted of a felony and whose DNA is in Philadelphia's database.

Joseph Szarka, the Philadelphia police crime-lab manager, said he was open to all possible strategies, including Morrissey's proposal.

"He has indeed offered the software to us at no cost, which clears our first hurdle to do this," Szarka said. "There are other open questions that I'm trying to address before we can proceed, like the legal issues."

State law, for example, does not directly address the use of databases to search for relatives.

Asplen said he saw nothing in Pennsylvania statutes that would prohibit familial searching. The FBI has no rules against such use of state data, he said.

Philadelphia would be limited to searching its local database unless it got cooperation from state police.

The state police sent a statement saying it would wait until there is a national policy before opening its database. "Although familial searching has the potential to be a great investigative tool, implementation at this early stage, without direct legislative approval and a standard national policy, is premature," Lt. Myra Taylor said.

Advocates say California, Colorado, Texas, and, most recently, Virginia did not wait. Last week, Virginia's director of forensic services agreed to take Colorado's software, which may be used to seek the so-called East Coast rapist and another man who raped and killed Virginia Tech student Morgan Harrington.

"There's no way lack of a national policy should prevent them from doing the right thing," Asplen said. "It's shocking that police could have this technology and not use it. Failure to move on this issue will cost people their lives."


Contact staff writer Faye Flam at 215-854-4977 or