Boston researchers said yesterday that police could identify many more rapists, murderers and other criminals by expanding the use of DNA databases to find close relatives of suspects whose DNA is recovered from a crime scene.
Today, law enforcement laboratories take crime scene DNA and compare it with large libraries of DNA obtained from convicts, looking for a perfect match. If there is only a near match, the researchers suggest, that person is likely to be a sibling, parent, or child of the suspect, providing detectives with a lead.
The technique has been used in a handful of cases, in which identifying a relative was the crucial break that allowed detectives to focus their investigation on a particular suspect and solve the crime. The new research, published in today's issue of the journal Science, concludes that familial searching could be a broadly effective tool for law enforcement. One computer simulation shows it would increase the number of times DNA would point to a suspect by 40 percent.
''This shows that familial searching is practical on a grand scale," said David Lazer, a coauthor of the article who is associate professor at Harvard University's John F. Kennedy School of Government.
But civil liberties watchdogs and other critics charge that the technique is flawed. Relatives usually share more DNA than strangers do, but not always. So investigators would have to check out false positives -- people who have nothing to do with the suspect, but whose DNA looks similar to the crime scene DNA by chance -- consuming time and invading the privacy of innocent people.
The technique also makes family members of criminals a part of investigations, even though they have themselves done nothing wrong. And critics say that members of minority groups and the poor would be unfairly targeted because they tend to be overrepresented in the DNA databases maintained by the FBI and state and local police.
''The genetic surveillance of innocents would be along racial lines," said Barry C. Scheck, a founder of the Innocence Project, which uses DNA to free people who have been wrongly convicted of crimes. ''I think it is a troublesome idea."
The study's authors -- Lazer, Frederick R. Bieber of Brigham and Women's Hospital, and Charles H. Brenner of the University of California at Berkeley -- stopped short of endorsing familial DNA searches, saying that ethical and other issues would need to be settled before a decision could be made on using the tool.
Law enforcement officials said yesterday that the technique could be useful, and the controversy surrounding it will be one of the main topics at the ''DNA Fingerprinting and Civil Liberties National Symposium" in Boston through tomorrow. The symposium is sponsored by the American Society of Law, Medicine, and Ethics.
Today, there are roughly 3 million DNA profiles in state and federal databases, according to the Science article. Since the early 1990s, searches of these databases have generated about 30,000 matches for investigators, but it is not known how many of these have led to convictions, according to Bieber. A Boston Police Department spokesman said its DNA database has generated 219 matches since going online in 2001, but it's not known how many led to convictions.
State databases typically begin with DNA fingerprints generated by analyzing blood samples of convicted felons, but there has been a movement nationally to expand the pool of people whose DNA is collected. Massachusetts has about 47,000 convicted felons in its database, according to Robert E. Pino, who administers the State Police database. Four states collect DNA from people who are arrested in feloy cases, even before they are convicted. New York lawmakers are considering entering anyone convicted of a felony or a misdemeanor.
Few states have rules about familial searching. In Massachusetts, the technique is legal, but State Police do not use it, according to Pino. He said that the technique could be useful, but that he would need more manpower to do the work.
Occasionally, Pino said, a near match will come up during the course of a search, and these leads are passed on to investigators, but systematically seeking out near matches and sorting through them would take more time.
In Virginia, the director of the state's Department of Forensic Science, Paul Ferrara, said that his office does not do familial searches, but that his staff sometimes comes across very close matches.
Once, he recalled, his office was convinced that a rapist must be the brother of a near match. He did not pass on the name to investigators, however, because Virginia has no policy addressing familial searches. ''It is very frustrating," he said.
The FBI does not use the technique, but it could be useful to investigators if legal, social, and other issues are resolved, said Joseph A. DiZinno, deputy assistant director of the FBI Laboratory in Quantico, Va.
DNA fingerprints are records of specified portions of a person's DNA. Given two similar fingerprints, a computer can calculcate the likelihood that the two people are related, but there will also be a chance that the two appear similar by chance. There is no threshold that would guarantee that two similar fingerprints come from related individuals.
People who have been convicted of crimes are more likely to have relatives that have been convicted of a crime, which would help the technique be successful. Bieber cited one study in which 30 percent of inmates surveyed said they had a brother who had also been in jail.
The study published today used a computer simulation to study how well the technique would work, if generally used. It used data randomly generated based on genetic principles, not information from real databases, meaning the results would have to be verified with real-world data.
A computer, the study said, could generate a list of near matches, ranked by how likely those people are to be related to the person whose DNA was collected at a crime scene. One possibility, the study's authors suggest, is to decide on an arbitrary cutoff, giving police access to only those partial matches that exceed a certain likelihood.
Law enforcement agencies today can't take full advantage of existing DNA matching procedures because they are overwhelmed with information. Across the country, Bieber and others said, there are large backlogs in processing DNA from crime scenes. One study in Britain, where police have been aggressively adding the DNA profiles of even minor criminals, showed that the main factor that drove more matches was not having more offender profiles in the system, but having more fingerprints from crime scenes to match against.
In addition, there are many cases in which laboratories have made matches, and disorganization or understaffing has meant that investigators never follow them up. In some cases, the same offender has gone on to rape or murder other people.
''It is disgraceful, to say the least," said Rockne Harmon, a senior deputy district attorney in Alameda County, Calif.
Critics also point out that law enforcement agencies keep the original DNA samples, a practice that could be abused if the government were to start using them to obtain other kinds of information.
Suzanne Smalley contributed to this report. Gareth Cook can be reached at firstname.lastname@example.org.