Slashdot Mirror
Twist on DNA Privacy
ConfusedVorlon writes "The BBC is reporting the conviction of a man for the murder of a prostitute 15 years ago. The interesting twist is that his DNA was not on record - 'But it partly match[ed] that of a youth's who was known to the police - but who had not been born at the time of the murder. The teenager, it turned out, was a close relative of [the murderor].' There has been concern in the past at the idea of keeping DNA of those interviewed but not charged with crimes. I haven't previously heard of the privacy implications of being related to a criminal/suspect. If you've done nothing wrong, you've nothing to fear?"
There are two ways of comparing DNA that are used. The first, more common method, is to look at something called Restriction Fragment Length Polymorphism (RFLP). The spaces between known regions that can be cut vary. Look at enough of these and you can statistically say with some confidence that its your person.
The second method, which isn't used as often last I saw, involves looking at the DNA sequences in regions which do vary. Perhaps well studied regions like those associated with organ rejection in transplantation.
Both methods are more reliable than the people carrying out the work. With RFLP, there is a certain amount of subjectivity involved in calling two sizes the same. With sequencing, contamination is easy.
I'm sure the procedures have improved since I last observed them. The weakest link in both procedures is the human factor.