Computational protein design and discovery

Protein design has traditionally relied on an expert's ability to assimilate a myriad of factors that together influence the stability and uniqueness of a protein structure. As many of these forces are subtle and their simultaneous optimization is a problem of great complexity, sophisticated sequence prediction algorithms have been developed to aid in the design of novel proteins by providing quantitative analysis of the sequence-structure relationship. This review discusses some of the major developments in computational protein design, focusing on common inputs to the calculation and several often used search methods. We also highlight accomplishments in computational protein design, ranging from simple core redesign of an existing protein, to the design of new functionalities (catalytic or ligand-binding), and finally to a large-scale design of de novo proteins.