[jgershen]

That's almost exactly the opposite of what this article says. To summarize it briefly, Vardi reminds us that the importance of proving P!=NP is to enhance our understanding of the underlying issues. Also, he devotes a significant fraction of the article to reiterating that worst-case complexity is not the same thing as tractability.

I'm not really familiar with protein structure prediction, but I don't think that it involves solving NP-complete problems. Scouring Wikipedia articles about protein structure prediction suggests that the techniques used are broad, varied, and distinct from the classical set of problems in NP. If someone's doing work in this area and is attempting to solve a problem in NP I'd love to hear it.