In 1984, Alec Jeffreys, a professor of genetics at the University of Leicester in England, was studying the evolution of genes. He was particularly intrigued by the intron--the series of bases in chromosomes sometimes referred to as junk DNA, whose functions are not clearly understood. When the genetic code is transferred to the complementary mRNA molecules prior to being converted into proteins, these intron sequences are snipped out by special editor enzymes.

While analyzing fragments of DNA, Jeffreys noticed that introns were often made up of the same sequence of bases repeated over and over again, and that these sequences varied dramatically from one individual to the next. Indeed, it is because of these highly variable regions that the DNA from two people never breaks down into precisely the same pattern.

Although Jeffreys had no special interest in murder cases at the time, it occurred to him that the individual patterns, displayed as a DNA fingerprint, could become a powerful forensic tool. Biological evidence left behind at a crime scene could be matched to the person who committed the crime. Blood, semen, saliva, hair roots, or bone--any material containing nucleated cells,--could provide a tissue sample. However, a method for obtaining the ìfingerprintî needed to be found.

The DNA fingerprinting technique Jeffreys devised depends on DNA probes--small fragments of single-stranded DNA produced by machines called DNA synthesizers. The probes bind (by complementary base pairing) to specific base sequences of DNA from samples taken at the scene of a crime and from suspected perpetrators. Several probes are used, each labeled with a radioactive tag. This allows fragment patterns from the samples to be visualized on film and compared.

A 1987 Florida sexual assault case marked the beginning of widespread acceptance of DNA evidence in criminal trials. In fact, there have been over 20,000 cases in which attorneys have framed arguments using data derived from DNA tests on human tissues. In addition to attorneys, there are others who have welcomed the chance to use the results gleaned from DNA testing. Incarcerated criminals, claiming to have been wrongly convicted, are attempting to reopen their cases, encourage by a well-publicized case involving a Canadian who had his decade-old rape conviction overturned.

As with comparison of blood types in paternity suits, the DNA test works best when it can be argued that a defendant's DNA does not match the DNA at the crime scene. In other words, a defendant can usually be ruled out easily as the perpetrator, but it is much more difficult to prove that a particular defendant, and no other person, could possibly have committed the crime. However, it is often the case in criminal trials that the probability of a DNA sample having come from any person other than the defendant is so infinitesimally small that a jury is readily sold on the defendant's guilt.

Two procedures for DNA fingerprinting, both used on Jeffreysí original technique, are currently in use. The more time-consuming analysis, called Restriction Fragment Length Polymorphism (RFLP), may take several weeks to complete, and it must be based on good DNA samples; that is, the samples must be relatively large and uncontaminated by DNA from other organisms. Furthermore, extreme environmental conditions can damage DNA and make standard RFLP analysis uninformative.

In RFLP analysis, minute DNA fragments are removed from the entire molecule by means of specific nucleic enzymes. Radioactively tagged DNA probes bind to the DNA areas that are used to establish identify, and X-ray film detects the radioactive pattern. The X-ray film is then developed, and the DNA fingerprint is compared to those made from other samples. Multiple samples are compared, seeking a minimum of four matches. Once a convincing number of matches has been found, statistical predictions are made as to how likely it is that two people at random in the whole population could have that degree of a pattern match.

The second test-- the Polymerase Chain Reaction PCR)--is quicker to complete, taking only a few days, and can be done with much smaller tissue samples. When time is of the essence in forensic evaluations, this is the test that is used. In some respects, it can be thought of as a molecular photocopying procedure. It uses repeated cycles to reproduce a target area of DNA until enough copies are available for analysis. Thus, PCR itself is not an analytic tool; rather, it facilitates forensic applications by allowing a scientist to take a sample of DNA, which would generally be insufficient to detect the characteristics of the DNA, and amplify it. Because the products generated by one sample can serve as templates in the next cycle, the number of amplified copies doubles with each cycle. Thus, 20 to 25 cycles of PCR potentially yield about a million fold reproduction.

Because PCR is not an evaluation of strictly random portions of the DNA molecule, statistical projections based on this test are not as convincing as those based on RFLP. For example, assessment of your DNA and mine using PCR might indicate that there was only one in a thousand chances that those two samples could have come from the same person. Using our samples in the RFLP test, the statistics would say that there was only one chance in a billion that the samples were from the same individual.