DNA conclusive yet still controversial, Carnegie Mellon professor says

Although the odds that DNA evidence found at a crime scene will match by chance the DNA of a person who was not there are infinitesimal, controversy continues about DNA identification and its use in criminal investigations, says Carnegie Mellon University Statistics Professor Kathryn Roeder. Roeder will present a historical overview of the use of DNA identification on Tuesday, April 25, during the Annual Symposia of the National Academy of Sciences in Washington, D.C.

Almost 28,000 cases nationwide have been prosecuted with help from the FBI's data bank of DNA profiles, while at least 170 people have seen their convictions overturned on appeal thanks to DNA evidence. Nonetheless, the use of DNA evidence in appeals has been impeded by political considerations and legal uncertainties, according to Roeder. "After all other legal avenues have been tried, the hope of any innocent person is that biological evidence from their cases still exists and can be subjected to DNA testing. But DNA's value to free the wrongfully convicted can be attained only if political leaders allow its full application," Roeder said. "Thousands currently await the evaluation of their cases."

In the early phases, technical disputes among scientists impeded the use of DNA evidence, Roeder said. One of the earliest controversies to erupt over DNA testing was the magnitude of genetic diversity among people of different ancestry.

Some controversy remains concerning the so-called "cold hit" technique, in which investigators search a DNA database to find a match of DNA found at a crime scene and then collect other evidence to build their case -- as opposed to first identifying a suspect through other evidence and then using DNA to confirm their case. Some critics claimed that this practice could snag an innocent person, but Roeder has demonstrated through her own research that the likelihood of a false hit are miniscule -- in one case, for example, it was about 1 in 26 quintillion, a probability so slight it needn't be shared with juries, Roeder said.

"The jury can't handle such small numbers. We would do them a service to simply tell them it matches or it doesn't match," Roeder said.

Roeder began her career as a biologist, and much of her current research is focused on using statistical tools to understand the workings of the human genome and the nature of inherited diseases. She is a member of the Bioinformatics and Statistics Genetics Group, which includes researchers in the departments of Statistics and Biological Sciences at Carnegie Mellon, and the departments of Psychiatry and Human Genetics at the University of Pittsburgh. The group's primary research goal is to develop statistical tools for finding associations between patterns of genetic variation and complex disease.

Source : Carnegie Mellon University