The misconception here is that the DNA evidence has a 1 in 85000 chance of getting the right man. Going from a vague recollection of Ben Goldacre's book "Bad Science", if the evidence is wrong then there is are 84999 other people the evidence points to.

I can't fully remember though.