How did Baker and his coworkers solve the problem? The key to success here was their initialdevelopment a few years earlier (1999) of the Rosetta computer program.18 It assembles shortstructural fragments from unrelated protein structures with similar local sequences in theProtein Data Bank and simultaneously optimizes sequence and structure with respect to thetarget backbone conformation. Monte Carlo optimization was used in the calculations with anenergy function that treated van der Waals interactions (6-12 Lennard-Jones potential),hydrogen bonding and solvation effects; sidechain orientations were sampled from a largelibrary of rotamers. The program generates many putative solutions and ranks them in terms of energies.
Most important, Rosetta was designed to be a general program both for protein structureprediction and design, and it has continuously been developed since its inception, with a large cadre of users and co-developers. The key idea to build proteins from short fragments can betraced back to the work of Jones and Thirup, who showed that in the context of automatedprotein model building into crystallographic electron density maps, assembling proteins fromknown substructures is an effective strategy.19Baker and colleagues went on to show that a wide range of protein structures could be designedusing the Rosetta software. While protein design was initially just focused on designingstructures, more recent work has aimed to also design advanced protein functions. This still posesmajor challenges in terms of understanding protein dynamics, structural transitions, allostery,catalytic effects, and so forth, and is thus an area of active research