DDG is an automated and hardened version of the latest DDG protocol (developed in the DiMaio lab (Park et al J. Chem. Theory Computation 2016 vol. 12 p. 6201) and further defined at Cyrus), designed to run stably at scale. We have benchmarked these methods extensively against experimental data and found them to be more accurate than competing tools, such as FEP, and to offer very high throughput due to the algorithmic efficiency of Rosetta. Additionally, Cyrus has made improvements to DDG which have significantly improved its predictive power.

Design w/ Sequence Based Symmetry includes all of the protein design and modeling tools in Rosetta, but with core modifications to allow the specification of arbitrary symmetry. For example, for a 3-fold symmetric trimeric (C3) viral surface protein, this would require that each monomer has the same sequence and structure as its neighbor. The original full algorithm description is here, “Modeling symmetric macromolecular structures in Rosetta3” DiMaio et al PLoS ONE vol. 6 e20450 (2011). There are multiple examples in the literature, such as “De novo design of a four-fold symmetric TIM-barrel protein with atomic-level accuracy”, Huang et al Nature Chem Bio vol. 12 p. 29. (2016).

Design w/ Flexible Backbone incorporates various degrees of backbone flexibility into protein design. These methods are commonly used for the identification of mutations likely to improve affinity at a protein/protein or protein/small-molecule interface, especially if the user plans to test the output of computations in a library format and desires more mutational combinations to test.

Davis et al (2006) Structure 14 265

Ollikainen et al (2015) PLoS Comput Biol 11(9)

Humphris et al (2008) Structure 16 1777

Smith et al (2008) J Mol Biol 380 742

Smith et al (2010) J Mol Biol 402 460

Smith & Kortemme (2011) PLoS One 6(7)

Protein stabilization will be accomplished using various design and optimization methodologies. One method that is often successful is the coupling of Design with a Position Specific Scoring Matrix (PSSM) to direct sample that have been observed in homologs of the parent structure (Lehmann et al (2000) Biochim Biophys Acta 1543 408), This approach is useful for reliably identifying mutations that can stabilize a protein, and is particularly successful when focused on buried residues (see, e.g. Korkegian et al (2005) Science 308 857 and Goldenzweig et al (2016) Mol Cell 63 337 and Lehmann et al, Biochem Biophys Acta vol. 1543 p. 408). This method will more exhaustively sample all mutations in the core in order to more broadly search for stabilizing mutations.