Fix issues with Gasteiger atom typing Riley and Doug noticed that the gasteiger typing didn't work in the presence of virtual atoms. This was a known issue, but this commit should fix that. Also add unit tests for Gasteiger typing on the cannonical amino acids and nucleic acids. This involves annotating the database params files with bond type and formal charge information, as well as putting in code to deal with connection points in the gasteiger typing. All the default loaded, non-patched full atom residue types should be cleanly and accurately(?) typed. No integration or unit test changes expected.

• Summary

Fix static CMake build Add System directive for external static cmake build h

• Summary

Reducing the number cycles used for loop modeling and docking for the antibody_h3 integration test. This should speed it up for debug integration tests. A possible problem with only using one iteration is that some things (e.g. constraints) may not be properly handled on iteration 2. To exercise these issues, a specific test should be used and should be tacked onto a test for something else, so I feel fine about this. The antibody_H3 test will change.

• Summary

making "Signed" RDF fingerprints In order for RDF fingerprints that represent things like hyrdogen bonding, product of charges, etc to make sense, we need to keep track of the "sign" of the function. Specifically, a case where partner A is a hydrogen bond donor and partner B is an acceptor is a fundamentally different signal than hte reverse case. So the RDF code has been rewritten to accomodate this. additionally, a few bugfixes to the ligand highres docker code hts_io integration test changes all unit tests pass

• Summary

Fixing NDDP and ODDP RosettaScripts movers

• Summary

Changes to the InterfaceFeatures reporter to make it work with MySQL MySQL requires text fields that are used as a primary key to have a maximum length, so I added lengths to these fields. I removed a foreign key constraint that failed in MySQL because it did not reference a unique row in the referenced table. I removed trailing whitespace. Expected integration test change: features_database_schema

• Summary

Reducing the number of inner cycles from 3 to 1 for loop modeling and centroid docking in the SnugDock integration test. This should ensure that it does not time out when run in debug mode.

• Summary

Split match integration test Split each of the three match integration tests commands into their own tests. This should help with the length of the debug-mode match integration test on the test server.

• Summary

Comment out a problematic unit test. KicClosureTests::test_with_cut_fold_tree works on my machine, but not on the mac.clang build on the test server. That worries me, but right now I don't know how to debug it so I'm just removed the test. This commit also fixes the warnings that were breaking the linux.gcc builds.

• Summary

Fix the unit test build. This also removes some debugging messages that were cosmetically changing a lot of integration tests. Don't know how I missed these in the first place.

• Summary

fixing a bug that affects the old way of calling hybridization protocol

• Summary

Publish my work on the loops refactoring project. One of the goals of the loops refactoring project was to eliminate the massive code duplication in protocols::loops::loop_movers. This commit accomplishes this goal, and will hopefully make it easier to develop new loop sampling algorithms. None of the refactored code is used by default. It can be enabled by passing these flags to the loopmodel application: -loops:remodel perturb_kic_refactor -loops:refine refine_kic_refactor Most of the new code can be found in protocols::loop_modeling and all of it has thorough API documentation. The new code performs as well as the existing code on the Kortemme lab 12-res loop benchmark, but it doesn't yet support all the existing code's features (e.g. NGK and some options that tune KIC). No integration test changes expected. Unit tests passed on my machine.

• Summary

Miscellaneous updates and bugfixes for density/xray/symmetry code. Make quiet option for symmdef file generation quieter. Missed some Py_DECREF calls in phenix interface code. Removing numeric deriv checks in model-map error code.

• Summary

[minor update of SandwichFeatures] - write resfile automatically (NOTAACFHWY for loop residues and surface heading strand residues) - default value of CB_b_factor_cutoff_for_electrostatic_interactions = 10,000 instead of 60 which is mentioned in http://spdbv.vital-it.ch/TheMolecularLevel/SPVTut/text/STut09aTN.html in order to count all possible electrostatic interactions [integration tests comment] - should not change other than features and features_database_schema (if any)

• Summary

Add per-residue controls to hybridize protocol

• Summary

Temporary #ifdef to fig bug with Rosetta crystal refinement. This will be removed once boost has been updated.

• Summary

Legend: