A thermodynamic definition of protein folds
Modern techniques in structural biology, like homology modeling, protein threading, protein fold classification, and homology detection have proven extremely useful. For example, they have provided us with evolutionary information about protein homology which has in some many cases lead directly to therapeutics. Due to the importance of these methods, augmenting or improving them may lead to significant advances in understanding proteins. These methods treat the high-resolution structure as a static entity upon which they operate, however we know that proteins are not static entities---they are polymers that exist in an enormous array of conformational states. Therefore, we propose to model the proteins from a statistical thermodynamic viewpoint based upon their average energetic properties. We show that this model can be used to (1) better characterize the partial unfolding process of proteins, and (2) reclassify the protein fold space from a new perspective.