New tools and apps:
New Movers, Filters, etc:
Updated or bugfixed classes and modules:
selector=option in a tag.
MutateResidue now stronger than a disulfide
Updates and bugfixes:
mhc_epitopescoreterm, implemented using MHCEpitopeEnergy, allows packer-compatible de-immunization of proteins using ProPred or pre-computed database epitope predictors. The latter can be generated for IEDB and NetMHCII databases using mhc-energy-tools. Local de-immunization can be performed with AddMHCEpitopeConstraintMover. Integration with nmer/NMerSVMEnergy will be implemented in the next release.
Tools to export InteractionGraph to external code, so that the packing step can be done with QUANTUM COMPUTERS OH MY GOD IT'S THE FUTURE
A disulfide optimization mover
RotamerSetsObjects framework ** PruneBuriedUnsatsOperation
S), three-letter (e.g.
DSE), basename (e.g.
DSER), or full name (e.g.
RelaxScriptManager, to ensure that relax scripts are read from disk once and once only, on first demand, and in a threadsafe manner. (The
RelaxScriptManagerhas no user-facing interface.)
extras=tensorflowoption during compilation) to facilitate development of machine learning-based protocols.
-output_ligands_as_separate_chains, causes Rosetta to reassign chain ID of a ligand when it shares with a peptide chain.
RosettaAntibodyDesign Application released - preparing for published paper. Pre-print available here: [https://www.biorxiv.org/content/early/2018/02/23/183350]
dock_glycans bug fixes
ShearMover bugfix. Yes, this was one of the very earliest Movers we wrote. Imagine our shear disbelief at discovering there has been a bug in it for most of a decade.
More resfile commands: We have 3 new resfile commands: CHARGED, AROMATIC, and PROPERTY. The third is a general command that takes any ResidueProperty. Currently, it only works for Cannonicals, but perhaps that could be generalized int the future for NCs.
ResfileCommandOperation - Applies the equivalent of a resfile line (without the resnums) to residues specified in a residue selector.
CreateSequenceMotifMover - Simple mover to Create a sequence motif in a region of protein using the SequenceMotifTaskOperation. Uses psueo-regular expressions to define the motif.
-alternate_3_letter_codes pdb_sugar. For more information, please see WorkingWithGlycans
The XML supporting RosettaScripts has been totally refactored. Rosetta now validates input XML files against an "XML Schema", and can better determine at the start of a run if all XML options are legal and functional. The XML reader can now pinpoint errors much more specifically and offers more helpful error messages.
You can now also get commandline help for XML-enabled classes with the -info flag; e.g.
-info PackRotamersMover and a blank, formatted template rosetta script by running the
rosetta_script application without giving the
A consequence of this refactoring is that pre-Rosetta3.8 XML scripts are usually no longer legal XML - we offer a tool to convert the old style, pseudo-XML into the current format at tools/xsd_xrw/rewrite_rosetta_script.py (this is in the tools toplevel folder, not the main code folder). The vast majority of classes have complete documentation; when you find one that does not, complain to (doc feedbacks email) and let us know!
A new Job Distributor, JD3, is ready for use. This is mostly invisible to end users, but will allow more complex protocols to be crafted instead of as multi-step and multi-app instructions. Look forward to cool JD3-enabled apps in future releases.
Although most improvements were in Rosetta3.7, we continue to improve the fraction of unmodified PDBs Rosetta can handle. (Don't worry - we've always been able to handle canonical protein well - but we are doing an ever-improving job with strange stuff like the GFP fluorophore, chemically concatenated ligands, glycans, RNA, etc).
Rosetta turned on Cxx11 features in its C++. This deprecates the compatibility of a lot of older compilers. See https://www.rosettacommons.org/docs/latest/build_documentation/Cxx11Support for more information.
We've tweaked the build system such that the built executables are named a little more simply. The tripartite names (rosetta_scripts.default.linuxgccrelease) will work as before, but now the two-part names rosetta_scripts.linuxgccrelease) will always point towards the default build, instead of the most recent build.
We have a new scorefunction brewing! It doesn't have its official name yet, but it is published. You can try it out with -beta_nov15 on command line and
<ScoreFunction name="beta_nov15" weights="beta_nov15"> in RosettaScripts in the meantime. Note that the -beta_nov15 commandline flag is necessary for any use of the new scorefunction, since certain scoring-related objects must be initialized differently (meaning that it is not currently possible to score with talaris2014 and beta_nov15 in the same session of Rosetta.)
Rosetta's Ramachandran scoring code has been greatly refactored. The software now supports Ramachandran potentials for arbitrary amino acids. These are lazily loaded, so they do not contribute to Rosetta's memory footprint unless they are needed. The refactored Ramachandran code is now part of the
rama_prepro score term in the beta_nov15 scorefunction. The Ramachandran scoring also now allows different Ramachandran potentials for positions preceding proline residues and for positions that do not precede proline residues.
The RosettaAntibody protocol has been streamlined and is closer to being a single application instead of a constellation of scripts. New paper outlining the functionality
RosettaScripts now supports inclusion of script fragments in other scripts
loops::restrict_kic_sampling_to_torsion_stringrepaired (for KIC loop modeling)
After releasing Rosetta 3.5 in 2013, Rosetta transitioned to a weekly release system. Rosetta 3.6 is weekly release v2016.13-dev58602. We are marking it as 3.6 to meet the needs of users that require numbered releases instead of weeklies. This large set of release notes collects changes released in weekly releases since their inception.
Since Rosetta 3.5, the default scorefunction was updated two generations, to Talaris2013 and then to Talaris2014 .
We’ve also updated the default minimizer to LBGFS, which is expected to provide better performance on the average problem.
Released Sunday, June 2, 2013
Released Sunday, March 25, 2012
Released Wednesday, August 31, 2011
Released Friday, March 11, 2011
Released Friday, March 11, 2011