Reduced models of proteins and their applications

Reduced computer modeling of proteins now has a history of about 30 years. In spite of the enormous increase in computing abilities, reduced models are still very important tools for theoretical studies of protein structure, dynamics and thermodynamics. Very simple, highly idealized lattice (and recently also off-lattice) models could be studied in great detail, providing valuable insight into the most general factors governing structure stability, folding kinetics and interactions responsible for characteristic two-state behavior near the folding temperature. More complex models now enable modeling of real proteins on the level of low to moderate resolution, allowing us to address more detailed questions. Ab initio protein structure predictions, still being far from a routine task, have become feasible. When supported by evolutionary information from multiple sequence alignments and potential local and/or global structural similarity to known structures, reduced modeling opens up new areas of comparative modeling, thereby complementing contemporary structural genomics.