Blood, Hair and Heredity

The murder of Nicole Simpson and Ronald Goldman seems tailor-made for DNA fingerprinting. There are no known witnesses to the two bloody killings. No murder weapon has been recovered. Detectives have not reported finding any standard fingerprints at the crime scene, and as a result the case against O. J. Simpson remains largely circumstantial. Those are exactly the sort of gaps that spurred British geneticist Alec Jeffreys to invent, in 1984, what has been called the magic bullet of criminal investigations. All a crime lab has to do is compare, say, the DNA of a murder victim to DNA in blood on a suspect's clothes. If the samples do not match, the suspect likely walks. If they do, DNA has fingered him.

Since it was first used in the United States in 1987, some 24,000 cases have involved DNA evidence. In 30 percent, it has cleared people mistakenly suspected of crimes. It has even freed from prison about a dozen men wrongly convicted of rapes or murders, and helped bring guilty verdicts or pleas in thousands of other cases. In April, Virginia carried out the nation's first execution of a murderer convicted largely through DNA fingerprints. The power of the prints comes from their precision, which is immensely greater than standard analyses of blood, semen or hair (chart). Police have reportedly tested blood at the murder scene and found that it includes O.J.'s type. No one has said whether it is A, B, AB or O, but at least 13 million other people share every blood type and those numbers are unlikely to sway a jury. But DNA tests could. Prosecutors will compare DNA in blood scraped from the crime scene to Simpson's DNA, and DNA in hair from a blue knit cap discovered at the murder scene to DNA in O.J.'s hair. If the judge allows it, the district attorney will also test blood found in Simpson's house and Bronco to Nicole's and Goldman's DNA.

The defense insists on running its own DNA tests -- for good reason. Although DNA fingerprinting has attained mythic status in the public mind -- one critic calls it "a fist on the scales of justice," outweighing all other evidence -- it isn't an infallible crime solver. It is problematic enough that the very term has been disallowed in some courts, replaced by "DNA typing" or "DNA profiling" so jurors are not bedazzled into thinking that DNA evidence is as unambiguous as an ordinary fingerprint. That's because DNA tests do not compare every one of the 3 billion chemical units of a person's heredity. (If they did there would be no dispute that a match is a match: DNA is unique to every individual except identical twins.) Instead, the tests compare only three to six regions of DNA.

Here's how it works: technicians extract DNA from blood, semen or hair cells. They dissolve it in enzymes that act like chemical scissors, snipping the DNA molecule into thousands of pieces, explains Mark Stolorow of Cellmark Diagnostics in Maryland. The lab then separates the pieces by size. The pieces are exposed to special radioactive probes that home in on specific DNA landmarks on the pieces. X-ray film reveals what size piece each landmark falls on. The whole process takes at least a month and results in a DNA fingerprint that looks like a blurry version of a grocery product's bar code, all lit up with radioactive probes at certain spots on each "bar" of the "code." Finally, the lab compares that fingerprint to the fingerprint of another sample -- say, blood on a murder suspect's clothing to blood from the victim. If they match, FBI and other experts claim, the odds are overwhelming that they came from one person: the probability of finding a match by chance is 100 million to 1.

If prosecutors find no incriminating DNA matches, the case against O.J. is dramatically weakened. Exculpatory DNA evidence has proved more powerful even than eyewitness testimony. Last year Walter (Tony) Snyder was released from a Virginia prison after serving seven years for a 1985 rape that new DNA evidence said he did not commit, even though the victim identified him at the trial. "DNA is often portrayed as a tool of the prosecution," says Keith Brown, president of the DNA testing lab GeneScreen in Dallas, "but it has done far more for defendants."

Even if DNA tests do turn up a match between, say, Goldman's blood and blood in Simpson's Bronco, it does not mean O.J. is as good as convicted. The defense can challenge DNA on several fronts:

The first hurdle for the prosecution will be to persuade a judge to allow DNA results as evidence. Four state high courts (Washington, Arizona, Vermont and Massachusetts) have ruled DNA evidence inadmissible until scientists agree on its precision. Twenty-two others have decided it can be introduced. California case law is contradictory. One of its appellate courts upheld the admissibility of DNA evidence and two threw it out, all in 1992. The fight over admitting DNA evidence will be one of the most important in the entire case.

Unlike labs that test for alcohol or drugs, DNA labs are not required to meet any standards. And the complicated science offers many chances to botch the job, as O.J.'s lawyers are sure to tell the jury. The first concern is contamination. "Say some material is recovered off a driveway," says Jonathan Koehler of the University of Texas. "When the lab is setting up the test, [the technician] accidentally dribbles some of the defendant's blood into . . . the driveway blood. You're going to get a match. It's happened." In one 1989 proficiency test, human errors occurred in two of 150 DNA samples. (Labs say the error rate is now much less.) That error rate is what jurors should hear, argues Koehler, not the million-to-1 number sometimes claimed for DNA's precision. Which of the competing figures a judge allows jurors to hear could strongly influence how persuasive they find any evidence.

If experts testify that there is only one chance in millions of an incriminating DNA fingerprint being wrong, the defense should pounce. The FBI calculates that kind of likelihood by extrapolating from the 3,000 DNA samples in its database. But defense testimony could undermine such assertions. In a 1991 paper, Harvard University geneticists Richard Lewontin and Daniel Hartl argued that in some population groups the chances of two people having the same DNA profile are higher: the huge odds against a false match could be grossly overstated. Other geneticists agree that the precision of DNA tests might vary by ethnic group, but they say it is hardly relevant. "There may be a question about whether [the chance of a false match] is 1 in 100,000 or 1 in a billion," says geneticist Victor McKusick of Johns Hopkins University, who headed a 1992 panel of the National Academy of Sciences that endorsed DNA fingerprinting. "It is a nonissue."

But for one jury it was very much an issue. In a 1992 California rape trial, prosecution witnesses put the chance of a false DNA match in a semen sample at 1 in 189 million; defense experts testified that it could be 1 in merely 65,000. The statistical dispute raised enough doubt that the defendant was convicted only of attempted rape -- and jurors said he would have been acquitted outright had he not lived next door to the victim. That's a lesson Simpson's defense team would do well to remember: even if they successfully sow doubt about the DNA tests, the jury will still be weighing other evidence, too.

How it Works

Fingerprint Experts compare the pattern of ridges and whorls.

Blood Serologists compare proteins and antigens that can

vary from person to person.

Hair A crime lab compares the color, shape and microscopic

characteristics.

DNA Technicians compare 3 to 6 regions of DNA obtained

from blood, semen or hair.

Theoretical Accuracy

Fingerprint Perfect: no two fingerprints are identical.

Blood Varies: 1 out of 2 people are type O. But only 1 in

1,000 share all 9 proteins typically tested.

Hair Poor: can identify race but not individuals.

DNA Hotly debated: only 1 in millions may have the same

patterns.

Admissibility

Fingerprint In all states

Blood In all states

Hair In all states

DNA Varies by state

Possible use in O.J. case

Fingerprint None have been reported.

Blood Samples from murder scene were tested.

Hair Tests run on hairs found in glove and caps.

DNA Samples from O.J.'s house, car and crime scene may be

tested.