Updating BoltzmannRotamerMover in protocols/simple_moves Adding a complete version of BoltzmannRotamerMover to protocols/simple_moves

• Summary

Merge pull request #71 from RosettaCommons/nollikai/BoltzmannRotamerMover Adding BoltzmannRotamerMover to protocols/simple_moves Created a new mover that calculates Boltzmann probabilities for each rotamer at a given position and then selects a rotamer based on these probabilities. If the position is designable, the mover will calculate a Boltzmann probability for each amino acid and then select an amino acid based on these probabilities. Boltzmann probabilities are calculated as p_i = e^(E_i / kT) / Z, where E_i is the change in energy of substituting the current rotamer with rotamer i, Z is the partition function (sum of Boltzmann factors for all rotamers), and kT can be set using set_temperature( Real ). The mover only changes one rotamer per apply. If randomize_resnum is set to true, the mover will select a random position out of the repackable and designable residues on which to perform the move. This position can be set using set_resnum( Size ). This mover solves a common problem I have encountered with traditional design, where multiple runs tend to output the same low energy sequence. Instead, the goal of this mover is to sample around local minima to obtain a large number of diverse low energy sequences. This is especially true when combined with local backbone moves, and I will commit another mover that performs coupled backrub and Boltzmann rotamer moves in the near future.

• Summary

In AtomTree mover added a tag that sets a special fold tree for antibodies I created a new integration test to test this called: Splice_min_fold_tree No other integration test changes expected.

• Summary

update to latest master

• Summary

try to reduce changes in swa_rna_gagu_04_clustering by forcing std::endl at end of a tracer output.

• Summary

a few small bugfixes for crystal refinement: cartesianMD now works; optionally specify minimum weight on xtal data

• Summary

Corrected a problem is CutChainMover that refrenced a residue at position total_residue+1.\n In splice I reverted changed done to find_nearest_res calls inside Splice::aplly

• Summary

corrected reference to non-existant residue

• Summary

Fixing bug in PyMolMover introduced in r42695. Coloring by energy and sending custom colors should now work as expected.

• Summary

Adding a funciton to ResidueGraph to regenerate the vertex and edge indices. This should be called when you are adding/removing nodes and edges to ResidueGraphs. Fixes bugs in smallest ring size detection algorithm. Added regenerate graph index to assign_gasteiger_atom_types(). This will regenerate graph indices (see above) before trying to assign atom types. Bug fix for proline patches

• Summary

Forgot to add constant_seed flag to the integartion tests, (which caused them to fail): splice_out_generic splice_out_longer splice_out_shorter_loop modified: tests/integration/tests/splice_out_generic/flags modified: tests/integration/tests/splice_out_longer/4JB9.pdb modified: tests/integration/tests/splice_out_longer/flags modified: tests/integration/tests/splice_out_shorter_loop/flags

• Summary

Merge pull request #69 from RosettaCommons/weitzner/stepwise_ccd_slowdown Weitzner/stepwise ccd slowdown: The CCD update resulted in a slowdown in some cases. Rhiju noticed that the Ramachadran data files were being read in from the disk multiple times, which, in the way it was written before, would happen when constructing a new CCDLoopClosureMover. This problem would then be exacerbated when CCDLoopClosureMover was created on the stack in a for loop. This revision fixes that by getting a COP to the Ramachandran potential stored in the ScoringManager. Additional changes are introduced to the Ramachandran2B potential to allow arbitrary conformations to be evaluated, and the RamaCheck2B class is updated to use this new interface. This means you can now use two-body Ramachandran checks during CCD closure! Finally, this change set updates StepWiseProteinCCD_Closer to have a CCDLoopClosureMoverOP as a data member instead of creating it on the stack to reduce the number of times "new" is called in that protocol. This should improve the runtime of integration tests that use CCDLoopClosureMover, and all of those tests will change due to the lack of tracer output associated with reading the Ramachandran data files. See http://benchmark.graylab.jhu.edu/revision?id=115&branch= for the full list of changed tests.

• Summary

Fixing some PyRosetta compilation errors on clang platform

• Summary

Add error checking on -out:levels Additional error checking on -out:levels settings such that we don't segfault for bad options (e.g. ones missing a colon, or options with more than one colon.) Also change behavior such that if a tracer is listed twice, the second time takes precedence over the first.

• Summary

Merge pull request #68 from RosettaCommons/JWLabonte/Atom_output This branch has made a number of (I think) improvements to the `show()` methods of several classes. In particular: - [x] Added `conformation/Atom.cc`. - [x] Added `conformation::Atom::show()` (and an `operator<<`). - [x] Renamed `chemical::Atom::print()` to `chemical::Atom::show()`. - [x] `Residue::show()` now calls `conformation::Atom::show()`. - [x] Added optional parameters to `Residue::show()` and `ResidueType::show()` to cause the latter to call `chemical::Atom::show()` for each atom. - [x] Expanded what is output by `chemical::Atom::show()`. I also cleaned up the Doxygen in the files. I also added `GasteigerAtomType`s to sugar residues and ouput the results in the carbohydrates integration test to both test these new `show()` modifications and to test Gasteiger typing of sugars. no unit test changes Only the carbohydrates integration test changes, as expected.

• Summary

Applied the follwoing changes: Updated the Splice* integration tests to test for newly introduced changes to Splice mover Corrected a problem with arguments passed to find_nearest_res in Splice mover. Before it only looked for nearest res in chain A, now looks in all chains. I Changed LoopLengthChange mover so it now respects the cut_chain input. meaning: if the input cut_chain is X then it will remain X after llc mover is apllied. Added source PDB 3UMT to the H3 PSSM database Added to CutChainMover a functions that finds a chainbreak between input residues

• Summary

Legend: