It may be convenient for the police, but the use of DNA as evidence is full of pitfalls.
Everyone's DNA profile should be included in a database, according to Robert Williamson and Rony Duncan in their Commentary “DNA testing for all” (ref 1.). This approach has much to recommend it to the police and the prosecution, but it will do little to increase the accuracy of any prosecution in terms of justice.
In Correspondence2, David Ehrenfeld invokes a 1907 detective story to comment on problems of classical fingerprinting still relevant today3. If such problems are still arising after nearly 100 years, how much more likely are problems with a universal DNA-profile database only a few years after the introduction of this technology?
One such problem concerns depositing DNA by touch when items are handled. It is possible to detect such handling successfully4, a fact already used in a criminal case in Canada5. Who can remember every item that he or she has touched for longer than one minute in the past week? When the knife you used in a restaurant is identified as a murder weapon, what is going to be your defence?
There are other major problems associated with a universal DNA-profile database. Would foreign visitors be profiled on entry? What happens if the person who matches the crime-scene DNA profile has an alibi at a distant place? The database proposal places a reverse onus on accused people to show their innocence, even when there is no other evidence to link them to the crime. What are the error rates for each step in the DNA-profiling procedure? Because the probability of a duplicate match is so low, bayesian theory suggests that when there is no other evidence, error is likely to be a higher probability than a correct match.
Although the safeguards Williamson and Duncan suggest are implementable in the United Kingdom, Australia and other industrialized countries, they are impossible in countries such as South Africa because the cost of independent forensic laboratories is prohibitive. In many countries, most sample collection is done by police officers, not scene-of-crime scientists, and a lack of resources makes independent testing impossible. Samples from suspects, even in the United Kingdom, are taken at police stations. Would independent DNA-testing stations be set up?
Finally, of course, computerized DNA databases are no more immune to hacking than any other database.
The criminal-justice system is weighted against the defendant in many countries. The defendant often uses a legal-aid lawyer with little or no knowledge of DNA profiling; he or she is in front of a judge with little or no experience of DNA evidence; the defence is only informed of any DNA evidence a couple of days before the trial starts, resulting in little or no chance to verify independently data presented to the court, let alone reprofile the sample.
I accept without reservation that DNA evidence is a very useful tool in criminal prosecutions. However, the police and forensic services in the United Kingdom have at times failed to obey the rules with regard to DNA evidence (see, for example, cases referred to the Court of Appeal Criminal Division, refs 6 and 7). Even now, the system has faults that need to be addressed before any future developments are considered.
Williamson, R. & Duncan, R. Nature 418, 585–586 (2002).
Ehrenfeld, D. Nature 418, 583 (2002).
Cyranoski, D. Nature 417, 676 (2002).
van Oorschot, R. A. H. & Jones, M. K. Nature 387, 767 (1997).
R. v. Xie  ABQB 478.
R. v. B  EWCA Crim. 42.
R. v. Weir  EWCA Crim. 43.