As lawyers in the O.J. Simpson murder trial prepare to debate the admissibility of DNA evidence, two Harvard population geneticists continue to challenge the validity of DNA fingerprinting.

Agassiz Professor of Zoology Richard C. Lewontin '50-'51 and Professor of Biology Daniel L. Hartl object to the forensic use of DNA finger-printing, or profiling. They contend that the accuracy claimed by current statistical techniques is overstated.

The O.J. Simpson case represents one of the most notable uses of DNA profiling, a technique used all over the country to convict criminals and clear innocent suspects. The statistical methods now being challenged by Harvard researchers are employed in thousands of cases each year.

If done correctly, DNA profiling has the potential to help unravel cases that would otherwise go unsolved.

Last Friday, Edward W. Honaker, a welder imprisoned for 10 years on a rape conviction, was freed after DNA profiling proved he could not have raped a 19 year old girl in 1984.

Honaker was represented by two lawyers from New York: Barry Scheck and Peter J. Neufeld. The Innocence Project, directed by Neufeld, has helped to free between 15 and 20 inmates around the country by using DNA evidence, according to the New York Times.

Simpson's defense has enlisted Scheck and Neufeld to help deal with the DNA evidence; however, the Los Angeles District Attorney's Office has Lisa Kahn on their staff. Kahn is a leading prosecution DNA expert.

Despite threats that he would bar DNA evidence because of the prosecution's delay in providing material to the defense, Judge Lance A. Ito has said that he will admit such evidence.

But "the battle is not over with" said Dr. Mark V. Bloom, assistant director of the DNA Learning Center at Cold Spring Harbor, New York. The defense will continue to fight against the DNA fingerprinting, because it could be some of the most incriminating evidence brought against Simpson.

The Science

DNA profiling is "a tool for exclusion," says George F. Sensabaugh, Jr., a professor of forensic science at the University of California at Berkeley. "If I am innocent and DNA evidence can prove it, I want testing done. [But], if I am guilty and DNA evidence will implicate me, I don't want it done."

"About 30 percent of suspects are excluded [from guilt] by DNA testing," says David H. Kaye, Regents professor at Arizona State University. "DNA profiling has resulted in many cases being resolved that would have gone unsolved."

The technique, also known as DNA typing or fingerprinting, is based on the fact that although the majority of DNA is the same in all humans, certain unique patterns exist in every person, except identical twins. DNA profiling distinguishes among samples by using these differences like a regular fingerprint. Experts have called the technique "the most significant breakthrough in forensic science since the development of fingerprinting."

The two methods of DNA profiling are restriction fragment length polymorphisms analysis (RFLP) and polymerase chain reaction analysis (PCR). RFLP, the more powerful test, analyzes specific locations of the DNA that differ widely across people (see graphics).

"RFLP methodology was first introduced and developed in England," says Paul B. Ferrara, director of the Virginia division of Forensic Sciences and chair of the American Society of Crime Laboratory Directors Accreditation Board. The test "gives rise to a virtual identification with a match at 5, 6, or 7 locations giving statistics showing us something like [only] one in 100 million" people could match.

"That method suffers, however, because it takes 8-10 weeks on average, requires a relatively large sample, and the DNA has to be in good shape," Ferrara says. "A relatively large sample would be a blood stain the size of a half-dollar or a seminal fluid spot about the size of a dime."