From Stan Brown, Forensic Science Northern Ireland
I was disappointed by your article “Between prison and freedom” (14 August, p 8 and p 5). I can state confidently that DNA profiling, regardless of public perceptions, is not “generally seen as infallible” by forensic science organisations.
DNA profiles may often be partial, or mixed (or possibly contaminated), but even obtaining a clear profile still only addresses the “source” question; “whose DNA is this likely to be?”
The second or “activity” question, “How and when did the DNA get to be on the exhibit or at the crime scene?” requires a detailed evaluation by Biology Reporting Officers. They are trained to estimate the appropriate “likelihood ratio” – the probability of the profile being found, given the prosecution’s proposition of events, versus the probability given the defence’s. While precise numbers may not always be possible, because of unquantifiable variables, if the evaluation cannot differentiate to a sufficient degree between the two propositions, it will state so in the court report.
In the UK, scientists have no incentive or vested interest in proving defendants guilty. All forensic DNA providers work independently of the police (unlike in the US) and all reports and evaluations are peer-reviewed.
Advertisement
Your statement that “little distinction is made, either in labs or in court, between samples of different quality: DNA is just DNA” is unfortunate. Any laboratory approaching DNA in such a cavalier way would fail its quality accreditation under ISO 17025 and forfeit permission to upload profiles to the National DNA database.
From Herzl Regev
Software packages to evaluate forensic results statistically should, as you say, be open source so that all labs can check them. Further, every test should be evaluated by each lab using several such packages, to avoid the possibility of a software bug skewing the results of all labs, leading to a wrongful conviction.
Beer Sheva, Israel
Carrickfergus, County Antrim, UK
