Over nearly two decades, a serial killer has shot and strangled at least 11 people, often dumping their battered bodies in alleyways of Inglewood and Los Angeles.

Most were black women or girls, the youngest just 14. The latest was found last year, shrouded in a garbage bag.

Police have determined through DNA and other evidence that the killings were the work of a single person. But the DNA does not match any of the millions of genetic profiles of convicted criminals in law enforcement databases, and detectives have few other clues.

Now Los Angeles Police Department investigators want to search the state's DNA database again -- not for exact matches but for any profiles similar enough to belong to a parent or sibling.

The hope is that one of those family members might lead detectives to the killer.

This strategy, pioneered in Britain, is poised to become an important crime-fighting tool in the United States. The Los Angeles case will mark the first major use of California's newly approved familial searching policy, the most far-reaching in the nation.

But the idea of scrutinizing families based exclusively on their possible genetic relationship to an unknown suspect makes privacy advocates and legal experts nervous. They argue that it effectively expands criminal databases to include every offender's relatives, a potentially unconstitutional intrusion.

"There is kind of a queasiness about having the sins of your father come back to haunt you," said Stanford University law professor Hank Greely, who supports familial searching despite those concerns. "It feels like we're holding people responsible for the crimes of their family."

Because the technology isn't perfect, families with no connection to the perpetrator inevitably will be investigated, some scientists and legal experts say.

The FBI and California law enforcement officials long resisted the approach, fearful of inciting legal opposition and a public backlash. They yielded only after aggressive lobbying by prosecutors, who pointed to some dramatic successes.

In North Carolina, for instance, investigators found a partial match to genetic evidence in a murder-rape case: a man who shared 16 genetic markers with the perpetrator. He couldn't be the killer -- he'd have to match on all 26 markers of a typical DNA profile -- but he could be a relative.

Investigators learned the man had a brother who had lived near the crime scene. They obtained DNA surreptitiously through a discarded cigarette. When tests linked the brother to the crime scene evidence, he confessed.

A person who had spent 18 years behind bars for the crime was freed.

Similar stories have come from Britain, where investigators have been using more sophisticated tools. In 158 searches since 2003, relatives have led authorities to the perpetrators in 18 cases that might never have been solved otherwise, British authorities said.

In 2006, British police were stymied in their hunt for the "shoe rapist" -- a man who collected the shoes of his many rape victims as trophies. DNA left at the crime scenes did not match anyone in Britain's national database.

But a woman whose DNA was in the database because of a drunk driving arrest had a similar genetic profile. She told authorities about a brother, a South Yorkshire businessman. In a search of his business, police found dozens of stiletto heels hidden behind a trap door.

In the U.S., officials are trying to balance the technology's promise and perils, pledging to use databases to identify relatives only for serious crimes and when all other leads have been exhausted.

"If you are going to do familial searches, I think it is important to really reserve them for just the most egregious crimes where there is an ongoing threat to society," said population geneticist Dan Krane, an expert on DNA evidence at Wayne State University in Detroit. But local labs, unrestrained by state and federal policies, are already starting to expand the technology beyond that limited use.

'Database creep'