There are a whole series of issues around DNA evidence. DNA isn't an absolute identifier. Like fingerprints, there's a misconception about just how accurate it is - most people think that fingerprints and DNA can tell you that unidentified Person A is definitely the same person as Suspect 1. Rather, we need to think of it as being able to tell you that Person A is definitely not Suspect 1. Looking for similarities in two potentially related patterns can be misleading, as it's possible then to ignore the differences and distort your conclusions. It's much more accurate to look for differences to rule people out.

Misidentification does happen, this isn't just a hypothetical situation.

Here in the UK, we have a massive DNA database, created and administered by the police. It has one in 10 of our white males listed, but four in 10 black males. There are also a scarily huge number of children on the database, as well as witnesses (i.e. people who were never even suspects), people who've given DNA samples in order to have themselves ruled out, people who have been acquitted, and those who were never charged. Once you're on the DNA database it's nigh on impossible to get off - your details will be kept for 100 years.

There's no doubt that DNA evidence can be useful indeed, but if you have a large database which you trawl for matches, it's possible for the police to find a match and then construct a case where there is none. We've seen this with fingerprints, and I'm sure that there are or will be examples based on DNA too, where innocent people find themselves in jail because their DNA happens to be similar to the person who actually committed the crime.

The problem with proposing a 'proving innocence' rule would be that the police would never go for that sort of restriction of their power, regardless of which country they are in. They see this as evidence - which is it, used properly - and any fruit of the poison tree proposal would never fly.

In Hunt's case, it seems from your description that the justice system was flawed in the first place - he was incarcerated despite contrary evidence. Plus overturning convictions always seems to be a bit of a problem, I suspect because it means that the authorities made a mistake and there's resistance from them to admit that. So is the problem here that there wasn't accurate enough enough DNA evidence in the first place, or that it was misused, or that the system itself was flawed?

The problem with having a big DNA database goes beyond the issue of whether the police correctly use the evidence to corroborate rather than as a central pillar of their case. Once you gather lots of data about people in one spot, all sorts of things become possible. Corruption, such as selling data illegally, becomes possible and again, we've seen this happen in the UK with the National Police Database. Identity theft is bad enough if your credit card details are nicked - but you can get new cards. You can't exactly get new DNA.

There's mission creep, where increasingly wide uses are authorised, such as racial profiling or the identification of men accused of not paying paternity support.

And there's the problem of the mass criminalisation of society - treating everyone as if they are a suspect undermines trust in the police and government, which has some pretty nasty knock-on effects in terms of damaging the fabric of society and the relationship between people, particularly young males, and the authorities.

Overall, I think that we (i.e. all countries using DNA evidence) need to have a clearer understanding of exactly what DNA evidence means, what it can and cannot do, the risks of using it, and under what circumstances it can mislead. We need to resist building big databases of DNA samples, no matter how attractive that seems. And we need to stress the importance of corroborating evidence - no one should be convicted by DNA alone.

If you're interested, the Open Rights Group (disclaimer: I started it!) in the UK has a lot of good stuff on its wiki about this and related issues: http://www.openrightsgroup.org/orgwiki.