In 1988, 20-year-old Lynette White was fatally stabbed in South Wales. The murder went unsolved for 15 years, until a fresh DNA sweep of her apartment in 2000 turned up spots of blood on a skirting board that had been missed the first time around.
British police ran the results through a national DNA database of known criminals, but didn't turn up anyone with an exact match. They did, however, notice someone whose DNA profile was close: a 14-year-old boy who was not even alive when White was murdered but who had gotten into trouble with the cops.
DNA testing of the boy's family eventually led police to Jeffrey Gafoor, the boy's paternal uncle, whose DNA exactly matched that of the blood sample. When questioned, Gafoor admitted to murdering White.
The case was a dramatic example of "kinship analysis," which could become more common as the practice of collecting DNA for crimes increases and the technique becomes more systemized and efficient, researchers said Thursday.
But widespread use of the technique raises issues about civil liberties violations, they caution.
Guilty by association
Studies have shown that a person's chances of committing a crime go up if a parent or sibling had previously done so. And a 1999 U.S. Department of Justice survey found that 46 percent of jail inmates had at least one close relative who had been incarcerated.
Britain has adopted a policy where almost any run-in with the law, even minor ones, will allow police to collect DNA. In the United States, the rules vary depending on the state. Currently, the U.S. criminal database contains DNA samples of about 3 million people.
Despite the potential usefulness of the technique, experts worry that maintaining a DNA database of criminal relatives could reflect — perhaps even amplify — demographic disparities already present in the criminal justice system.
"Right now, if one looks at who is in the [DNA] database, it tilts heavily towards African Americans, Hispanics and to people with lower incomes," said study team member David Lazer, an associate professor of public policy at Harvard University. "Now, if you're looking at their immediate relatives, they'll be more likely to be African American, Hispanic and poor. It's going to shine the spotlight brighter and brighter on a certain corner of our society and not so much on other corners."
One solution, the researchers say in a paper published on the journal Science's Web site, would be to create universal DNA databases containing samples from every citizen, so that everyone is represented. Calls for such an action, however, have so far been rejected.
Another option: Use the kinship analysis for only the most serious cases, and set statistical thresholds so that innocent people are not needlessly harassed.
Any policy governing the use of kinship analysis will have to maintain a tricky balance between collective security and individual privacy, the researchers contend.
"On one hand, it's not right to put in a whole class of people who have never been convicted, arrested or suspected of a crime under lifelong genetic surveillance," Lazer told LiveScience. "But on the other, it would be morally repugnant not to catch a murderer if all it takes is a click of a button to activate the search algorithm."