Saturday, February 23, 2008

Gene police

The WSJ (Gautam Naik, A1) ran an interesting story on Britain's use of DNA samples in solving crime. KG at Crusader Rabbit beat them to it. Apparently, law enforcement in the UK use "familial searching" to narrow down possible candidates for a crime at which a DNA sample was collected. Sounds legit and beneficial, no? There's a but, though; it lies in a comparison between the US's and the UK's methods of collection...

The method, briefly:
In doing a standard genetic match, U.K. scientists look at 20 alleles [forms of genes] along a string of DNA. If a match is found for all 20 alleles for a pair of DNA profiles, there's virtually no doubt that both DNA samples come from the same person -- unless the samples come from identical twins. Near matches of only 17, 18 or 19 alleles may indicate a close blood relative.
As of now, the US collects DNA only (overwhelmingly) from convicted criminals and, if from not-convicted individuals, destroys the sample if a person is acquitted. This works out to about 2% of the population (5.6 million profiles). The UK, though, keeps every sample it gets, coming out to about 8% (4.2 million in England/Wales). Here (emphasis mine):
Anyone arrested -- including for minor offenses -- must provide a DNA sample, which stays in the database permanently, even if the person is acquitted. About a quarter of the profiles are of minors, some as young as 10.
The kicker, which worries me:
The U.S. is now considering following Britain's lead...

[Eleven] states have passed laws to allow DNA samples to be taken from people who have been arrested but may not have been convicted. The Department of Justice separately is finalizing rules permitting genetic profiles to be collected from anyone arrested by federal authorities. And in March, the FBI will host a symposium to discuss the merits of familial searching. One speaker is Tony Lake, chairman of the U.K.'s national DNA board.
Conflict of interest? No? Granted, there's at least one benefit:
A familial search looks for a near match, rather than an identical match. A near match may indicate that the sample in the database is that of a close blood relative, who can then help lead police to the actual perpetrator.
DNA profiling brings up other concerns as well, mostly ethical:

Civil-liberties groups oppose the rapid expansion of DNA databases, arguing that they risk placing sensitive personal information in the hands of the government...Leaked data, for example, could be used to deny insurance coverage or employment to people identified as being at risk for a genetic disease.

Forensic experts say such fears are overblown because databases hold only a tiny fraction of a person's DNA makeup. The information is stored as a series of numbers converted from a physical DNA sample.

Critics also say some research techniques pose ethical problems. They say a DNA-based "ethnic inference" test can provide uncertain predictions about the race of potential suspects that may mislead the police or reinforce existing prejudices...

...Alec Jeffreys, who invented the technique of genetic profiling in 1984, has questioned the need to retain DNA profiles of innocent people. He says the DNA database was originally set up only for convicted criminals...

A law as flexible as this worries me as well:
...Police were initially required [in 1995] to destroy DNA samples and profiles if a suspect's charges were dropped or if the person was acquitted. But that requirement was eliminated in 2001. Three years later, police were allowed to take DNA from anyone they arrested.
And is this nitpicking? You tell me whether it sounds like a nanny state or not:
Now, U.K. police are seeking powers to let them take DNA samples from people who are charged with minor infractions like speeding or littering...A senior appellate judge, as well as former prime minister Tony Blair, have called for the database to ultimately include DNA profiles of every U.K. resident.
Granted, they have successfully used familial searching to convict about 20 criminals. Does the end really justify the means?

No comments: