Washington DC

US campaigners against capital punishment are hoping that modern DNA tests in old cases will undermine public confidence in the death penalty.

In the next few weeks, genetic testing of an old sample could show that Roger Keith Coleman — executed in 1992 for rape and murder — was not guilty. The tests are being conducted at Canada's Centre of Forensic Sciences in Toronto, and were requested by the state of Virginia's governor, Mark Warner. If Coleman is exonerated, it will be the first time that genetic evidence is used to prove that the US death penalty has killed an innocent person.

Coleman was convicted in Virginia for the 1981 murder of his sister-in-law. The sample being tested was gathered in 1990 as part of an unsuccessful attempt to clear his name. In ordering the recent tests, Warner — a Democrat widely expected to run for president in 2008 — called Coleman's case “an extraordinarily unique circumstance, where technology has advanced significantly and can be applied in the case of someone who consistently maintained his innocence until execution”.

Only once before in the United States has DNA been tested in a case where the prisoner had already been executed. The tests, performed in 2000, were inconclusive about the guilt of the Georgia man, Ellis Felker, who had been executed four years earlier.

Retesting DNA samples is the only morally acceptable course.

But modern DNA tests have successfully exonerated some Virginia prisoners. Between 2002 and 2004, three people were cleared after tests were done on old biological samples, saved by a methodical employee in the era before such techniques were available. As a result, Warner asked for tests on a random 10% of all the samples that had been saved. Out of thousands of cases, there were 29 in which current DNA tests could have shown guilt or innocence. Two of the 29 were shown to be innocent and their convictions overturned.

Innocents at risk: the United States is in a furore over questionable DNA evidence. Credit: TEK IMAGE/SPL

Warner has now announced that the rest of the samples will also be tested, calling this “the only morally acceptable course”. Ellen Qualls, the governor's communications director, says that a rough extrapolation from the test results suggests there may be 20 or 30 exonerations from around 300 pertinent cases — an error rate that some would argue is unacceptable for a state that allows capital punishment.

The US public has long been divided over the death penalty. Each state decides whether to administer it or not: 38 states allow people to be sentenced to death, and Illinois has suspended executions while it examines the issue. Proof that innocents have died would certainly strengthen arguments against the penalty.

The issue of wrongful convictions does not apply just to cases that pre-date DNA testing. A growing scandal in Texas serves as a reminder that genetic tests are only as reliable as those carrying them out. On 4 January, a damning report was published on the Houston Police Department's crime laboratory. The DNA unit of the Texas lab was closed in 2002, after a series of exposés by local television news station KHOU. Now, an independent investigation, which has looked at 1,100 cases so far, charges the lab with problems including contamination of samples, misrepresentation of statistics in court and failure to use controls. In two cases, one of which put a man on death row, a clear absence of the suspect's genetic markers in a sample was called “inconclusive” and not reported.

Investigators blamed lack of funding and inept management, among other factors. “The complete lack of outside scrutiny of the crime lab's operations, procedures, and reporting of results allowed serious deficiencies...to become so egregious that analysts in the lab simply had no perspective on how bad their practices were,” the report reads. DNA analysed at the lab is being retested, and at least one person has been freed.

Nearly 95% of US police crime laboratories are accredited by a board that maintains national standards. Ralph Keaton, head of the accreditation project, says that Houston's experience is the exception rather than the rule. “From the beginning,” he says, “DNA evidence has been carefully scrutinized because it is so powerful.”