Why do they fold the way they do is still unsolved. Most of the studies are empirical.

Protein folding and unfolding is still not properly described by atomic scale simulation. Replica exchange molecular dynamics method somehow helps to explore free energy landscape of protein but it also computationally much expensive.

The ab initio prediction of the 3 D structure of even a small (say 100 residues) protein molecule given the sequence and environmental conditions (temperature,pH,solvent composition,pressure etc) is an unsolved problem even when one forgets QM and models everything in

the classical Newton picture.Due to combinatorial explosion of the many body system involved,the inadequate treatment of Coulomb forces,the polar,associative nature of water and the fact that thermodynamic quantities like free energies and entropies are not Newtonian averages over

the trajectory inspite of some interesting new theoretical tricks(Replica Exchange MD,Jercynski identity) powerful computers and numerical methods are absolutely necessary but not sufficient.One needs new natural intelligence work to develop efficient approximate solutions but such work is not properly supported and in fact hindered by legions of ambitious computers " über alles" adepts who by and large,ignore the

GIGO (Garbage In Garbage Out) principle of blind computations.

The missing part is to show (propose) an exact pathway of the protein folding/misfolding. The protein structure, hydration and dynamics in each step. No matter with what agent you are perturbing the protein structure (temp, PH, ions, denaturents, macro-molecules or may be stabilizing agents). Obviously one can do thermodynamic and spectroscopic study along with some computer simulation study for microscopic view.