The Effectiveness of DNA Fingerprinting
by Lamise Kassem

DNA identification is the latest in a long line of controversial forensic science methods (fingerprints, lie detectors, voice print). More than the others, however, this advance stimulates the imagination in ways akin to the mythical theft of fire from the gods. DNA-based techniques address growing fears about violent crime. They also respond to even more basic hungers; they seem capable of vanquishing the age old foe of uncertainty itself.

For many years now, scientists have argued about the reliability of forensic DNA fingerprinting which is considered to be a powerful tool, some kind of technology that was able to link blood, semen, or hair left at the scene of crime to a suspect's DNA. The question is, " Is DNA fingerprinting effective?" Experts have fought in so many courtrooms and have argued about this new technology. A flow diagram shows the process of DNA fingerprinting(refer to Bioscience, Vol.43, no.3, March 1993) A question arises about the result of this process and that is, "Are we getting the right (1) result ?" Many lawyers tried to find something to act upon this new technology that would make it an unusable tool in predicting who is the suspect.

The lawyers attacked the scientists on how they could calculate the odds that two people can have the same DNA profile. A DNA fingerprint is not unique, as say, a fingerprint. There could turn out to be two unrelated people to have the same set of alleles at the studied sites. The scientists were able to resolve a problem like this by calculating the probability by using the multiplication rule. The multiplication rule is basically where they estimate the frequency with which each allele occurs in the general population and then they multiply that frequency to possess the odds of the match (Scientific American, October 1994, pg.34). Before performing this calculation, it must be assumed that the alleles of different markers are inherited independently of one another.

After doing the multiplication, is it likely that DNA profiles match? If we arrive at a probable result, then the strength of this evidence is usually measured by a "match probability" the probability that an unrelated individual chosen randomly from an appropriate population will match the crime profile. ( Nature, Vol.368, 24 March 1994, pg.285). We must also take into consideration that even if we have a small match (2) probability that does not in itself, establish guilt. We can see the results of this tool by examining the examples below. In the first example, the circumstances are real, however, the DNA fingerprint evidence allowed investigators to distinguish between suspects : who were the father and his ten year old son. What basically happened was that a brother and sister were stabbed and the sister was also raped. She died as a result of the attack but the brother even though he was in a critical condition, recovered. The father was accused formally for the rape and murder of his daughter and the attempted murder of his son. The father was questioned and his defense was that his son had raped the sister and was stabbed by her during the struggle. Conventional serological testing of the semen stains may or may not unequivocally distinguish between father and son.

The DNA pattern from the father matched that of the semen evidence, whereas the DNA pattern of the son did not. Even though we might say since the father and son are related and they share some of the markers, it turns out that there are genetic differences that were specific to each family member. These differences allowed each piece of evidence to be unambiguously attributed to a specific individual. The father subsequently was found guilty and sentenced to life in prison. The results in this case highlight the power of DNA (3) fingerprinting, and the legal issues raised during the prosecution of this case are typical of those surrounding DNA-based identification technology as it is currently used today.

Another example of the utility of DNA fingerprinting is where evidence is left at the scene of crime. The evidence may be examined for the purpose of identifying the person or persons who might have committed the crime. It might seem that DNA analysis is most efficient in sexual assault crimes. In this crime, the biological evidence(e.g semen collected from vaginal swabs) and the location of that evidence is a convincing argument for the guilt of a suspect whose DNA matches the DNA results. Even more important is the often overlooked exclusionary or nonmatch result. Generally, if there is not a match between the evidence and suspect, a charge of a sexual assault becomes difficult to support in most circumstances. When charges are dropped as a result of DNA nonmatch, there is no forum for the examination of the results as there is in a match and subsequent trial.

A figure (Bioscience, Vol.43, No.3, March 1993, pg.150) shoes a result of DNA matching according to the samples that were studied. DNA samples were taken from three known individuals, which are shown on the autoradiograph : the victim, suspect 1, suspect 2, and from an additional control sample. From the autoradiograph, DNA samples on the lane (4) labeled "evidence" match the lane labeled "suspect 1". From looking at such results, the whole process is relatively easy to evaluate. There maybe many individuals who if not for the DNA evidence, would be as likely to be the guilty person as the defendant. If there were 500,000 such individuals, for example in a large city, the probability that the defendant is innocent after taking the DNA evidence into account would be at least one third (Nature, Vol.368, 24 March 1994, pg. 285). The reasons for the power of DNA-based identification system depends on two facts :

1). LOW INCIDENCE PROBABILITIES
2). MULTIPLICATION OF INDEPENDENT OBSERVATIONS

Even though it turns out that DNA fingerprinting is a powerful tool, its high efficiency and reliability of its results must not lead to over-interpretations of their meaning. Positive results are still based on probabilities and statistical analyses which always provide information about uncertainties rather than about facts. Also when the results reveal that one in a billion or in a trillion, this does not mean anything since this process deals with long calculations, which may not be justified. Also, in addition to DNA fingerprinting, more evidence is needed to reach the verdict of either innocence or guilt. (5) The result of all of this has showed that DNA-based identification is very important and has already been built in the forensic community and has become a viable part of the judicial system both in the United States and abroad. The Federal Bureau of Investigations performs 2,500 tests a year for federal, state and local prosecutors, tests that have helped to convict or exonerate tens of thousands of suspects.(6)

References

1). Identity Testing in Forensic Science Biosciencs, Vol.43, no.3, March 1993 Kevin C. McElfresh, Debbie Vining Forde, and Ivan Balazs.
2). How convincing is DNA Evidence ? Nature, Vol.368, 24 March 1994 David J. Balding and Peter Donnelly.
3). High Profile Scientific American, October 1994 John Horgan.
4). DNA Fingerprinting : Academy Reports Science, Vol.256, 17 April 1992 Leslie Roberts.
5). Science in Court : A Culture Clash Science, Vol.256, 7 August 1992 Leslie Roberts.
6). Statistical Evaluation of Dna Fingerprinting : A Critique of NRC's Report Science, Vol.259, 5 February 1993 Devlin, Neil Risch, Kathryn Roeder.