Miami — Tommy Lee Andrews has twice been convicted of rape in Orlando this year even though neither of his victims got a good look at him. The key evidence: his genes.
Now under appeal, the Andrews case is likely to provide the highest court ruling yet on the legal validity of a new test for identifying human evidence: DNA fingerprinting.
The process is a lab test that portrays the genes on a human chromosome like a supermarket bar code. From a hair, a drop of blood, or any other bit of biological evidence, a DNA pattern can be compared to that of a suspect much as the whorls of fingerprints are matched.
In another rape case approaching trial in Palm Beach County, a judge ruled last week to admit DNA fingerprint evidence indicating that the baby of a rape victim was fathered by the defendant.
On the other hand, rape charges were dropped against Tony Harris recently because, although the victim identified him, his DNA did not match that of the semen sample.
``The DNA really tipped the scales,'' says his attorney Jack Nants, ``because the rest of the case was balanced.'' Otherwise, a jury would have been left to choose between Mr. Harris's alibi and the victim's identification.
DNA fingerprints are not definitive.
Sometimes the tests produce no results. British scientists found last week that the old evidence was too degraded to shed new light on the rape conviction of Gary Dotson, who was identified by a victim who recanted years later.
The certainty of a DNA match depends on how rare a DNA type is and how well-preserved the sample. A test produced by a California company, Cetus Corporation, can work on very old samples with degraded, broken DNA fragments. But the results may only narrow gene types to those held by one in five people.
On better samples, the test by Lifecodes of New York can narrow the DNA type to 1 to 10 million - for even the most common gene types - according to Michael Baird, manager of forensic and paternity testing.
The scientific credibility of the tests make it tough to fight in court. ``We could not find a single scientist who thought it didn't work,'' says Hal Urhig, Tommy Lee Andrews' attorney.
But Mr. Urhig questions the theory behind the high certainties offered by scientists. While geneticists treat each band of DNA like a random variable, Urhig notes that genetic traits often go together - such as blond hair and blue eyes - and their DNA codes may not be as random or independent from each other as they are made out to be.
He also notes that Lifecodes projects extremely high certainties mathematically but their sample base is only about 1,000 DNA tests.
Ed Imwinkelried, law professor at the University of California, Davis, also questions some of the extremely high certainties that DNA testers offer, such as a 1 in 3 billion chance that a DNA match is mistaken.
``I have not seen any clear, detailed explanation of these numbers,'' he says.
Russell Higuchi, associate scientist at Cetus, also cautions that experts should be humble in their claims for DNA matching. ``It's done by people,'' he says, ``and people make mistakes.''
The best an expert can really say, he suggests, is that ``this has to be the person, unless we made a mistake.'' Attorneys can then attack or defend the quality control of the test.
DNA matching has been used in investigations nationwide. So far, trial courts have admitted DNA matching as evidence in four states: Florida, New York, Oklahoma, and Pennsylvania. The Andrews cases are the first to rise to the appellate court level.
On the civil front, vast numbers of paternity disputes are being diverted from the courts altogether as DNA matches determine who did or did not father a child.
It will take years, however, before DNA fingerprinting becomes as valuable as traditional fingerprinting as an investigative tool, Dr. Imwinkelried notes. Without a national library of individuals' DNA patterns, he says, DNA testing is limited to cases where suspects are known.
Few attorneys expect to be able to keep DNA evidence out of court. But in far greater numbers, DNA tests are keeping cases out of court. Either suspects plead guilty when faced with a test linking them to a crime, or, as in the Tony Harris case in Orlando, charges are dropped when the tests clear the suspect.
Traditional sleuth work is still required. In an Oklahoma missing persons case, a DNA test identified a drop of the victim's dried blood on the suspect's vacuum cleaner. But a jury acquitted the defendant because investigators never produced a body.