A woman claims one man is the father of her child. But he says he's never met her.

In a classic mystery, Sherlock Holmes or Perry Mason might have solved this paternity suit by matching eye color, or tracing social patterns. But they might have solved more cases if they only knew what students in a new class are learning.

The new course, offered by Medical School Associate Professor of Pathology Dr. Frederick R. Bieber, a medical geneticist at the Brigham and Women's Hospital, explores the scientific principles behind controversial techniques such as DNA typing, also known as DNA fingerprinting, and drug screening, which are introduced into the courts as evidence every year.

"As scientists, we act as advocates for the truth," says Bieber, who teaches Biology E-190, "Principles of Forensic Science," offered through the Extension School. "Attorneys must naturally be adversarial, because they either win or lose."

One of the most controversial topics covered in the course is "DNA typing," a method to identify a person on the basis of genetic information.

"It is a misconception that forensic genetics is only used to restrict freedom and convict people," says Bieber, who says that courses on this increasingly important field are rare. "Its use often results in exoneration of those falsely accused or incarcerated."

For example, he says, such techniques are used in paternity suits, inheritance cases and even in the military.

The actual science behind gene typing iscomplex, and highly studied. But the scientificcommunity is currently debating whether DNAevidence is accurate enough to be admitted incourts.

Gene typing can be thought of as functioning inthe reverse way of physical fingerprinting, thatis, to eliminate what appear to be matches. Ingene typing, DNA samples isolated from tissues arecompared to information in a computer databank todetermine how rare a particular gene is in thegeneral population.

A so-called "match frequency" describes thelikelihood that the sample could have come fromany random individual. As match frequencyincreases, the identification of a suspect becomesless certain.

Professor of Biology Daniel L. Hartl is onescientist who has problems with the use of DNAtyping as evidence. In an influential 1991 articlein Science which he co-wrote with AgassizProfessor of Zoology Richard C. Lewontin '50,Hartl argues that DNA typing currently has twomajor flaws.

First, he argues, the current DNA databank istoo small, and second, it is unclear how muchattention should be paid to ethnicity whenmatching samples.

"Gene frequency differs from subpopulation tosubpopulation," says Hartl. "When you startcomparing with a databank containing DNA types ofall ethnic groups [put together], the matchfrequency seems smaller than it actually is."

"[DNA typing] is so good a method for gettingscientific evidence that we shouldn't fudge thestatistics," he says.

The National Research Council (NRC), in a bookcalled DNA Technology in Forensic Science,proposed a "ceiling method" to get around theproblems with the process. According to NRCguidelines, any sample which matches the generalpopulation's gene pool more than 10 percent shouldnot be considered a conclusive match.

"The ceiling principle, as advocated in the NRCreport, makes a reasonable attempt to solve theproblem of population substructuring," saysBieber. "However, it must be stated that thisprinciple has no basis in mathematical theory orpractice."