Folding proteins: finding a needle in a haystack

I revisit the Levinthal paradox of protein folding kinetics. Efficient computational methods are now finding or approaching native states on model free energy landscapes for small molecules and chains up to about 60 monomers. Also, recent experiments show counterexamples to the 'thermodynamic hypothesis', i.e. they indicate that the global free energy minimum may not always identify biologically active conformations of proteins.