|In court, scientists map a murder weapon|
|By Kate Dalke
January 24, 2003
The case was tragic and bizarre. In 1998, a Louisiana doctor was convicted of attempting to murder his ex-girlfriend, a nurse. The murder weapon was a syringe of HIV-infected blood drawn from a patient under the doctor's care.
The doctor, Richard J. Schmidt, is now serving a 50-year prison sentence for the crime. His conviction hinged on genetic evidence that the HIV used in the attack had originally come from Schmidt's patient. Now, scientists in the case have described their analysis of the evidencethe first of its kind presented in a US criminal court.
The story, according to prosecutors, went like this: Schmidt arrived at his former girlfriend's house the night of August 4, 1994, and gave her what he said was a vitamin injection. Six months later, when the victim tested positive for HIV, she accused the doctor of infecting her. Circumstantial evidence indicated that Schmidt had injected the victim with blood drawn at his office from a male patient infected with HIV.
DNA fingerprinting has been widely used to link human suspects to crimes. This trial was different, however, because the virus mutates so rapidly that in a matter of hours its DNA fingerprint can change significantly.
Both the defense and the prosecution enlisted scientists to determine the relatedness of the two HIV strainsthe victim's and the patient's. A close match between the viral DNA sequences could help prove that the doctor had committed the crime.
The case marks the first time this type of evolutionary analysis, called phylogenetics, has been admitted as evidence in a US criminal court. Scientists for the prosecution said they found strong evidence that HIV had been transmitted from patient to victim. Scientists for the defense disagreed.
The jury sided with the prosecution and convicted Schmidt of attempted second-degree murder. The defense appealed to the Louisiana Supreme Court and then to the US Supreme Court. In spring 2002, the US Supreme Court rejected the appeal and the guilty verdict stood.
Now, scientists for the prosecution have published their account of the materials, methods, and analysis they used during the criminal investigation. Michael Metzker of Baylor College of Medicine in Houston led the project.
The researchers compared the DNA of the virus in blood samples from the victim, patient, and HIV-infected individuals in Louisiana. The idea was to see if viral DNA from the victim and patient were more closely related to each other than to that of others in the Lafayette area.
Specifically, they studied a gene that makes the outer coat, or envelope, of HIV, and found the most similarity between the patient and victim samples.
"The envelope genes clearly showed a relationship between the victim and patient," says David Hillis of the University of Texas in Austin, who testified on behalf of the prosecution. He reviewed the analysis, done by two independent laboratories, one at Baylor and the other at the University of Michigan, Ann Arbor.
The scientists also asked which way the virus had been transmitted. Unfortunately, the envelope gene used to show relatedness mutates so rapidly that the details of its evolutionary history were lost.
So the scientists constructed evolutionary trees for a gene that evolves more slowly. The victim's sequences were "embedded" within the patient's sequences, suggesting the virus could only have gone from the patient to the victim.
"You could clearly see that the victim's sequences were derived from within variations in the patient's sequences," says Hillis.
Looking at two HIV genes was important because they have different functions and mutate at different rates, adds Metzker.
Bette Korber of Los Alamos National Laboratory in New Mexico, who testified for the defense, could not be reached for comment.
The case is novel in some respects, but this is not the first time HIV transmission has made headlines. In 1991, the case of the "Florida dentist" shocked the public and health care workers alike when it was revealed that a doctor had transmitted HIV to his patients during their dental visits.
Andrew Leigh Brown, a researcher at the University of Edinburgh's Centre for HIV Research in Scotland, has studied phylogenetics for over fifteen years. In the mid-1990s, he helped investigate a small outbreak of HIV in a prison outside Glasgow, Scotland. The researchers found that viruses from 13 of 14 inmates were closely related and had been transmitted over the past few months. But they never linked the outbreak to an individual.
When doing phylogenetic analyses, it is easier to exclude samples than to make a positive match between samples, Leigh Brown says. "In DNA fingerprinting, you have much greater confidence that there's a real match."
The Schmidt case carried the added burden of presenting complicated scientific analysis to lawyers, the judge and jury. The analysis often included statistics and probability rather than cold, hard facts.
"The big challenge was taking a complex set of procedures and making them understandable to a jury of non-experts in Louisiana," says Hillis.
Keith Crandall of Brigham Young University in Utah works on cases of HIV transmission, and, as a graduate student in Hillis' laboratory, he worked on the Florida dentist case. In these types of analyses, he notes, scientists are "trying to infer the history of a virus."
"All we can do is analyze the data and put a probability on it. A jury has a hard time thinking about probability," he says. "Even I have hard time explaining what I do to my parents."
Hillis continues to receive inquires from prosecuting attorneys and convicted prisoners about this type of analysis. "I have no doubt that we will see more of these cases in the coming years, involving HIV as well as other pathogens."
. . .