The New York Times reports on the decision to rule out Myriad's BRCA1/2 gene patents.
On a gut level, I agree with the direction of the decision, but the finer legal parsing for this patent ruling might not hold up on appeal. The essential problem at stake in this ruling is how to interpret the patentability criteria to allow for patenting things like natural product drugs without allowing for patents on simple restriction fragments or PCR products of DNA and DNA sequences.
Although ideologically it seems nice and clean to simply say, "you can't patent natural products, only uses of them" (i.e. "method" patents), this idea is hard to implement in practice, since it creates all sorts of legal grey areas. For one thing, Europe doesn't allow method patents. In addition, though, method patents are probably difficult to enforce, in that one can patent A to do B, and someone else can sell A to do C, knowing that many people will use it for B. The entire point of the patent gets undercut, and all this just makes for more lawsuits and headaches all around.
Thus, there is the idea of the "transformation" of natural products as a way to allow for patenting isolated natural products, when such products are sufficiently different from their "native" form that there is something essentially different about their properties. As you can imagine, this isn't the most straightforward of criteria, hence the controversial nature of the new ruling, part of which hinges on a new interpretation of this criterion.
Update: As always, Ars Technica has an excellent summary of the decision. John Timmer manages to convey how closely the ruling hinges on the fine wordings of the patents and laws in question.