Giving science the hard cell

Since the advent of forensic DNA testing less than a decade ago, DNA profiling has become an integral part of a number of areas of the legal system.

DNA profiling is an umbrella term which covers a growing range of practical technologies, but as complex as these techniques may appear, DNA testing in any of its forms can fulfil only two basic functions – identity testing and relationship testing.

The applications of identity testing are most obvious and most widely reported, namely in forensic identification in sexual or violent crime, where a suspect may be identified by traces of blood or semen left at the scene of a crime. However, it would be wrong to believe that only criminal lawyers have to wrestle with the scientific basis of DNA evidence.

The use of DNA profiling for relationship testing extends to areas such as paternity testing in maintenance, inheritance or custody cases and to immigration casework, where entrance to the UK is sought on the basis of biological kinship to UK residents. Identity testing may also be used to resolve insurance claims – blood stains on the chassis of an ultra high-performance car were used to ascertain whether the driver was actually the insured party.

DNA profiling data may be of interest to practitioners of several branches of the law. The precise application of the technology in a given case may largely determine the type of DNA profiling undertaken. Relationship testing, for example, traditionally uses a technique known as multi-locus DNA testing, which results in the generation of the barcode-like pattern commonly associated with DNA testing.

In cases where foetal samples are being analysed, however, possibly in a paternity or rape case, single-locus testing may be employed, due to the tissue being mixed with maternal tissue. Similarly, in forensic cases where single-locus testing is the norm due to the limiting amount of DNA available, polymerase chain reaction (PCR) based DNA testing may be undertaken in cases where only trace amounts of DNA are available. Hence the nature of both the case and the samples available determine the type of DNA analysis undertaken.

Given that there is a range of casework where DNA testing is appropriate, and a wide range of tests available, questions regarding the value and reliability of such data must be raised.

While it is fair to say that the standards of technical practice and quality control in most UK DNA testing labs are admirably high, this does not mean that DNA evidence should be accepted at face value. A number of areas should be examined in detail before the DNA evidence and its interpretation by the original testing officer should be accepted. Some of these issues are complex and not suitable for lengthy discussion in this article, but they may be summarised (and simplified) under the following headings:

– Continuity of samples

The processing of biological samples to extract the DNA is a fairly laborious task usually involving the simultaneous treatment of numerous specimens, which after the first stage, all look identical. The unequivocal identification of any given sample through not only the DNA extraction stages, but all ensuing stages of the practical technique is a challenge in itself. Some labs address this problem by having two operators verify every sample transfer, while others use duplicate samples to ensure continuity. The possibility of misidentification is particularly worrying in cases where there are several suspects, but in any case this possibility must be considered and the probability of its occurrence evaluated. If there is any doubt at all, then a new test on the suspect's blood sample must be undertaken.

– Quality of DNA evidence

The analysis of all DNA evidence is, to a certain extent, subjective; this is particularly so with respect to the analysis of multi-locus data, where the definition of bands may be poor. It is essential that the analysis of the reporting officer be verified, as unscored or erroneously scored bands can significantly alter the implications of the data. It is also important that phenomena such as band shifting are properly and adequately considered in the interpretation and statistical weighting of the evidence.

– Statistical evaluation

Once DNA profiles have been deemed to match, then a statistical weight is attached to the data. In single-locus testing this necessitates reference to a database of frequencies of occurrence of the observed genetic traits in the population as a whole. Debate continues as to the validity of several aspects of this approach, but one of the most pertinent concerns is whether the database used can reasonably be applied to the suspect. People from geographically, socially or religiously isolated groups are far more likely to share their

genetic characteristics with other members of their peer group, so that calculating the frequency of occurrence of their genetic profile in the population at large is meaningless.

There are, therefore, a number of areas where an independent expert evaluation of DNA evidence can shed new light on a case, either by an examination of the issues above or by simply offering alternative interpretation of the data. DNA evidence, while undoubtedly reliable in general terms, is not infallible in every case, and given the weight that this kind of evidence often carries, it is vital that it is validated by an independent scientist.

Dr Karen Sullivan is a lecturer in molecular genetics at the University of Abertay Dundee and an independent DNA profiling consultant.