In an unusual attempted-murder trial, prosecutors in Lafayette, Louisiana, hope to use a new kind of genetic evidence to help prove that physician Richard J. Schmidt tried to kill his former lover by injecting her with HIV-infected blood. The trial, set for later in the spring, may include evidence from phylogenetic analysis, a technique that compares DNA samples from various sources to see how closely they are related. The prosecutors maintain that the HIV strain infecting the alleged victim, intensive-care nurse Janice Trahan Roberts, is most likely to have come from one of Schmidt's patients.
Phylogenetic analysis has never before been used in a criminal trial in the United States. Schmidt's lawyers have challenged its admissibility, arguing that the technique is far more uncertain than DNA fingerprinting, which is widely accepted in the courts. They also argue that in this case the controls used and the lab work are seriously flawed. The defense lost the first round, when Louisiana District Judge Durwood Conque ruled in January that the evidence could be admitted. But on 3 March, the defense filed an appeal, a move that may delay the trial--currently scheduled for 12 May--until Louisiana's third circuit court can rule on Conque's decision.
Trahan claims that when she tried to end a 10-year affair with Schmidt in August 1994, the doctor, who had been giving her vitamin shots, came to her house and gave her an injection against her wishes. In December, Trahan began having suspicious symptoms. A month later, she found out she carried the virus, and in May 1995, she accused Schmidt of deliberately infecting her. Schmidt has pleaded not guilty; his lawyers say he was at home with his wife on the night in question. As part of their investigation, the police obtained samples of blood from Trahan and Schmidt's only HIV-positive patient. That's where phylogenetic analysis comes in: A scientist at Baylor College of Medicine in Houston compared DNA from Trahan's strain to the one infecting Schmidt's patient as well as to viral sequences from 30 randomly chosen HIV patients in the Lafayette area and to hundreds of HIV sequences in the national database.
More than Schmidt's innocence or guilt may hang on the outcome. His case is a testing ground for a molecular technique that is likely to find plenty of forensic applications in the future, says Gerald Myers, an expert on HIV strains at Los Alamos National Laboratory in New Mexico. Besides its use in AIDS cases, phylogenetic analysis may eventually be used to trace other infectious agents--in cases of food poisoning or even biological warfare. "The horse is out of the barn," Myers says. "[This technology] is here, and cases like this are going to continue to happen."