Latest News

Detectives failed to find a suspect in tests comparing the material against a DNA database of convicted criminals. But the search yielded a near match to a convict with an unusually similar DNA profile. The convict may not have been the robber, but perhaps one of his relatives was.

The process of turning crime-scene DNA into a family tree of possible leads has been quietly undertaken in more than two dozen cases in New York City since 2009, when a divided state committee voted to allow the release of so-called partial match DNA data to investigators.

Three years in, the street-level experience in New York has been decidedly mixed: Law enforcement officials said they knew of no cases solved in New York City because of a lead generated by a partial match.

The reason, investigators suggest, may be because New York — unlike states like California, Colorado and Virginia — has not allowed authorities to use existing software that explicitly searches state databases for kinship connections in specific investigations. Instead, during routine searches for exact matches in the databases, detectives and prosecutors are notified each time a piece of evidence matches most but not all of a database profile, suggesting a possible relationship.

In the case of the Queens robbery in 2010, detectives could not request a family search be performed; they were merely informed that a DNA sample partially matched a profile in the database. Detectives chose to pursue the potential lead, developing a family tree from the DNA sample; that led them to secretly securing a genetic sample from the father of a convict. But after tests ruled him out, they concluded the robber in Queens had no blood connection to the family under scrutiny, said Phil T. Pulaski, the chief of detectives, at a public hearing.

It is not clear whether the more tailored search software used in other states would have ruled out the family in this case before any detective work was done. "It is incredibly inefficient because you can't target a particular case," Melissa Mourges, the chief of forensic sciences at the Manhattan district attorney's office, said of the process in New York. "Therefore a lot of work gets done upfront in cases that we would not ultimately pursue."

Chief Pulaski explained that New York City detectives had used searches of the DNA database to pursue potential kinship leads in 25 cases, but had hit dead ends in four of them: the Queens robbery, a 1996 rape in a Manhattan hospital, a 2005 murder in south Queens and a 2007 Bronx murder. Chief Pulaski described those cases in detail to the DNA subcommittee of the State Commission on Forensic Science last year. (A fifth case was closed after the victim recanted, Chief Pulaski said.)

The four cases described by Chief Pulaski offer a unique glimpse at new genealogical procedures being honed by detectives. "In this family tree, we have 17 family members," Chief Pulaski said, discussing the hospital rape case. "We have ancestors, siblings, nephews," he continued. The police took DNA from two brothers and two nephews of the convict, he said, as well as from autopsy samples from two relatives.

But none of that matched the evidence from the rape case. "We're at a point where we don't believe there is a biological relationship," Chief Pulaski said.

He did not publicly discuss the 20 cases in which detectives are actively looking at relatives and could still result in leads.

Before the policy change in 2009, medical examiners were permitted to share only the names of offenders whose DNA appeared to exactly match crime-scene DNA. When medical examiners saw a partial match — suggestive of a familial relationship — they kept it to themselves. The 2009 change allowed the medical examiner's office to notify investigators, inquiring whether they wanted to know the name.

"It's up to them to do the legwork and see if they can find relatives that might be appropriate to test," said Marie Samples, an assistant director of the New York City medical examiner's office.

Wary of setting off a civil liberties debate that has so far been mostly avoided, law enforcement officials in New York have been reluctant to call for using familial search software, content to experiment with the results from partial matches.

"It's not like we were jumping for joy that this was going to be the elixir," said Joseph L. Giacalone, a retired commander of the Bronx cold case squad. "But you have a suspect where in the past you wouldn't have anybody."

But in roughly two-thirds of the cases, New York detectives and prosecutors have declined to learn the names. Of more than 140 names generated from partial matches, Chief Pulaski said, the police declined in 93 cases.

Asked in what circumstances the police might forgo the information, Deputy Inspector Emanuel J. Katranakis told the subcommittee that in some of these instances there had already been "a conviction and an individual is serving a sentence."

Chief Pulaski added, however, that the police were "very careful to use this and consider this as an exoneration tool, as well as an investigative tool to see who did it."

Critics worry the authorities have too much discretion to use — or ignore — partial-match information, which routinely brings innocent family members under suspicion. "It's a directive to the police to use your imagination," said Robert Perry of the New York Civil Liberties Union. In addition, the group argues, minority communities may disproportionately fall under suspicion because the underlying DNA database, based on convicted felons, contains a greater percentage of black and Hispanic offenders.

Chief Pulaski said the police usually declined information about partial matches because of skepticism that the evidence was relevant to a case. At crime scenes, the police often collect cigarettes or beer bottles for testing even though passers-by may have left them.

Chief Pulaski described a 2007 murder in a Bronx park in which the police had found a baseball cap some 60 feet away from bullet casings.

Testing of the cap showed similarities to the DNA of a convicted offender, prompting detectives to learn more about three of his male relatives — his father and his two sons. Witness descriptions of the shooter, however, excluded the father as too old and the sons as too young, Chief Pulaski said, leading detectives to conclude the baseball cap had no connection to the killing.

Some here have looked to the expanding number of states where more highly targeted searches for family members have provided breaks in high-profile cold cases, notably the 2010 arrest of a suspect in the so-called Grim Sleeper case.

In that case, California law enforcement officials identified a Los Angeles serial killer, whose crimes stretched back to the 1980s, after a 2010 database search turned up a possible relation. Detectives gathered DNA from that man's father, which matched the old crime scenes; he was charged with the killings.

Such highly tailored software searches, so-called familial searches, are believed to detect family relationships better — particularly those among siblings — than partial-match results generated in New York.

"Being able to pursue partial match is important," said Mitch Morrissey, the district attorney in Denver and a leading advocate of familial database searching. "It does work sometimes, but if you're going to be doing this type of work, you really need familial search software."

But, he cautioned, the results are still just one step in solving a hard-to-close case. "The science gets you to a viable lead, but then you have to do a conventional investigation," he said.