Rosetta Web Servers

The Rosetta macromolecular modeling suite has a very wide range of applications, and although installing and running Rosetta, becomes increasingly easier over the years, some non-experts would still prefer to run it over a comfortable to use web-server. We have assembled here the list of Rosetta applications for which such a web-server exists. Bookmark for future modeling tasks.


Robetta (Baker Lab)
This is the original and oldest Rosetta server, created to do what Rosetta was created to do: structure prediction. Nowadays the server offers more than just structure prediction and includes the following functions. Robetta is known for its good performance in CASP competitions, though the human Rosetta operators usually get better scores. There are 4 main categories you can use Robetta for:

• Structure prediction: Given a simple fasta sequence of your protein, performs ab-initio (or homology based) structure prediction. Beware – this takes a lot of time. It’s worth to note that one can also use “Rosetta NMR” through this server by including NMR derived constraints into the folding simulations.
• Fragment libraries creation: Given a fasta file – creates PDB based fragment libraries for all overlapping fragments of sizes 3 and 9. (This is required for several flavors of local Rosetta runs).
• Interface Alanine scanning: Given the structure of a protein-protein complex, performs computational alanine scan over all interface residues (or on residues specified by user). Provides estimated ddG values for these mutations and allows the detection of hotspot residues. (By Tanja Kortemme)
• DNA-interface scan: Given the structure of a protein-DNA complex, this server estimates the changes to DNA binding affinity and specificity that result from exhaustive point mutations at individual amino acid positions in the interface.

RosettaDesign (Kuhlman Lab)
Rosetta design can be used to identify sequences compatible with a given protein backbone, or to detect mutations at protein-protein interfaces that would increase binding affinity. Some of Rosetta design’s successes include the design of a novel protein fold, redesign of an existing protein for greater stability, increased binding affinity between two proteins, and the design of novel enzymes.

RosettaDock (Gray Lab)
The RosettaDock protein-protein docking server predicts the structure of protein complexes given the structures of the individual components and an approximate binding orientation. The server uses the Rosetta 2.1 protein structure modeling suite. It’s important to note that this is a “Local docking” server, i.e. the input complex should be in the approximate binding site. RosettaDock has a module that can handle “Global docking” (i.e. with no information on the binding site) but this is too computational demanding for a web-server. (You can detect initial potential binding sites using protein interface predictors).

RosettaAntibody (Gray Lab)
Antibody modeling server for prediction of antibody Fv structures. The antibody homology modeling server is built on the Rosetta structure prediction suite comprising of knowledge-based techniques for template selection, grafting for non-H3 CDR loops, de novo loop modeling for creating CDR-H3, and docking to optimize the relative orientation of the light and heavy chains.

RosettaBackrub (Kortemme Lab)

Flexible backbone protein structure modeling and design server based on the “backrub” method (first described by Davis et al). The server input is a protein structure (a single protein or a protein-protein complex) uploaded by the user and a choice of parameters and modeling method: prediction of point mutant structures, creation of conformational ensembles given the input protein structure and flexible backbone design.

• Point mutations: This function utilizes the backrub protocol and applies it to the neighborhood of a mutated amino acid residue to model conformational changes in this region.
• Backrub ensembles:
  • Backrub Conformational Ensemble: Backrub is applied to the entire input structure to generate a flexible backbone ensemble of modeled protein conformations.
  • Backrub Ensemble Design: This method first creates an ensemble of structures to model protein flexibility. In a second step, the generated protein structures are used to predict an ensemble of low-energy sequences consistent with the input structures, using computational design implemented in Rosetta. The output is a sequence profile of this family of structures.
• Interface sequence plasticity prediction: First, the backrub algorithm is applied to the uploaded protein-protein complex to generate a flexible backbone conformational ensemble of the entire complex. Then, for each of the resulting structure, residue positions given by the user (up to 10) are subjected to protein design. All generated sequences are ranked by their Rosetta force field score. Sequences with favorable scores both for the total protein complex and the interaction interface are used to build a sequence profile.

RosettaFlexPepDock (Furman Lab)
Rosetta FlexPepDock is a high-resolution peptide docking (refinement) protocol, implemented within the Rosetta framework. The input for this server is a PDB file of a complex between a protein receptor (first chain) and an estimated conformation for a peptide (second chain). FlexPepDock was shown to be able to accurately refine the peptide structure starting from up to 5.5A RMSD of the native conformation, allowing full flexibility to the peptide and side-chain flexibility to the receptor.

FunHunt Server (Furman Lab)
FunHunt is a classifier of correct protein-protein complex orientations. The input to FunHunt are two possible orientations of a complex. A local docking run is performed on the two complexes using RosettaDock. FunHunt then uses features gathered from these docking runs – representing the local energy landscapes of the orientations, and chooses the near-native orientation among both (assuming that one of the orientations is the near native one).