A DNA recipe for injustice

Mixed results

A national DNA database could well increase, not reduce, the number of wrongful convictions

Three killers have been convicted this month of appalling crimes. Two were implicated through DNA profiling. The aftermath has seen renewed calls for both a universal DNA database and the death penalty, bolstered by a belief that advances in scientific evidence would eliminate the risk of wrongful convictions. Keeping everyone’s DNA profile in the universal database could make a significant impact on the investigation of crime. But there are also cogent reasons why it would increase the risk of convicting innocent people.

The UK database is already the largest in the world. Profiles produced from crime scenes are now easily checked against that database, which at first mostly included only those convicted of serious offences. Now the net is much wider, with people who have been neither charged nor convicted still on the database. As a scientist who works on behalf of the defence in assessing DNA evidence – and challenged the validation of low copy number DNA in the Omagh bomb trial – I have some experience of the difficulties involved in such cases.

DNA profiling technology has improved from days when we needed a splash of blood or a good-sized stain. We can now obtain profiles from something that may have been only touched, and match DNA from just a few cells. From a system that had match probabilities of thousands, we are in the era of billions. This is where the problems begin. For many cases, the issue is not: “Is it my DNA?” It is: “How did it get there?”

The police are likely to conclude on the basis of a database match that they, more likely than not, are investigating the guilty party. The investigative process can then follow a pattern of accumulating circumstantial evidence. Once this process results in a charge, an error can be compounded by the jury failing to appreciate the difference between threads of evidence all dependent on that first “identification” by DNA. Juries, police, and lawyers can struggle with the many misunderstandings of probabilities, conditional match probabilities, odds and frequencies.

The police claim prosecutions will not be mounted on DNA evidence alone. This is not correct. There are cases where DNA alone could provide compelling evidence for a prosecution (for instance, a full profile obtained from an intimate sample from a rape victim when the defendant claims no contact) and cases where the DNA is irrelevant. There are no hard and fast rules.

Nevertheless, an “innocent” match being used to assemble a circumstantial case in court raises the very real prospect of false convictions. Excluding any DNA evidence from the case if it has been used to identify the suspect in the first instance might protect against this.

I suspect most people imagine that all DNA profiles obtained emanate from blood, semen, or some other clear stain. But often there is no visible stain at all, and the profile is a mixture. Mixtures create the potential for more difficulty. By way of illustration: if I have profile AB and you have profile CD, our mixed cells would have a profile ABCD. However, the same profile could be produced by two people with profiles AC and BD, or AD and BC. If this mixture was found at a crime scene, we now have six “suspect” profiles. If the person with the BD profile is unlucky enough to live in the area where the crime was committed, BD now needs to explain why he has no association with the material found at the scene. In fact, a mixed profile could generate about 60,000 suspects.

Technical and procedural issues associated with this powerful criminal investigative tool have been, and are being, resolved. However, the difficulties will be compounded and amplified if more individuals are on the database. The larger it becomes, the greater the chance is of a fortuitous “hit”, false conviction, and unnecessary stress on individuals and resource deployment by the police. Success, in thinking that we are convicting more criminals, could encourage further steps that will increase the errors.

Failure to use DNA properly will almost certainly result in unintentional but significant damage that could threaten its more restricted but beneficial use. This is a wide-ranging and important debate that requires an understanding of the scientific and procedural aspects of criminal justice systems and the frailties of both.