ABSTRACT
With the recent advances in the eld of computational protein design, we have seen many successful examples of de novo designs, including novel protein-protein and protein-DNA interactions, as well as enzymes capable of catalyzing novel reactions. However, the rate of success remains low, which suggests that our understanding of the sequence-structure-function relationship is far from complete. There are many formidable challenges that need to be overcome before we can reliably design new protein molecules at will. Here we review three topics that will be important in the future progress of computational protein design: backbone exibility, negative design, and experimental approaches.
