In other DNA related news, Scalia's dissenting opinion regarding DNA sampling in Maryland v King was my favorite thing I read this week.
There's some interesting math behind the practice of using DNA matching as sole proof in criminal cases. The stats are normally presented to the jury as though it was a one in five million chance the person is innocent...but if the size of DNA databases starts to grow, that could lead to several hits. Additionally, the stats do not factor in the chance that the sample was contaminated, or the chance that your DNA ended up somewhere randomly rather than intentionally.
End message: it shouldn't be treated as perfect.