commit f26e6cb0d8ed7edfe2157071420eb181ef491e06 Merge: 2ad4dae 8acb6c7 Date: Wed Aug 26 08:56:07 2015 -0700 Merge pull request #737 from RosettaCommons/rhiju/fix_match_test_instability fixing Matcher test instability Sweet, now integration.release_debug (linux) and integration (mac) versions of match tests actually match, including order of output of HIS vs. HIS_D, etc. (except for a nan vs. -nan). commit 2ad4daebf1ac800a61dd9fcd9d0ef54cf41f49ba Merge: db49f81 4fd6c33 Date: Tue Aug 25 18:00:03 2015 -0400 Merge pull request #680 from RosettaCommons/everyday847/chiral_flip_patch Generate D-amino acids using patches commit db49f814cfeb3b328d7ea0e8ca1705cb06dddabe Merge: 8b31a73 8d72b79 Date: Tue Aug 25 11:33:28 2015 -0700 Merge pull request #740 from RosettaCommons/calebgeniesse/cmake-dev Fix CMake Mac graphics build commit 8b31a732470048c2b2c693a621e79b460ff13ff9 Date: Tue Aug 25 14:03:33 2015 -0400 Adding mac.clang.ninja_graphics test commit a19fa6a1a60627a94d59211b50d0db3c0ab90326 Merge: f70cef6 fa30ee0 Date: Tue Aug 25 00:34:01 2015 -0700 Merge pull request #739 from RosettaCommons/vmullig/fix_swm_tests Trying to fix failing stepwise Monte Carlo integration tests. I think I've tracked the problem down to a point at which remove_variant_type_from_residue was being passed a residue and a pose that didn't contain that residue. This was an existing issue in protocols/stepwise/modeler/util.cc that was only revealed by my recent changes. Fixed now. commit f70cef625f09eaaf0dea21c68da40c6fee94db83 Merge: f386b3c ac4aff4 Date: Mon Aug 24 14:37:00 2015 -0700 Merge pull request #738 from RosettaCommons/vmullig/require_semicolon Removing semi-colons from some macros to remedy beautifier issue Removing semi-colons from the runtime_assert, runtime_assert_msg, and runtime_assert_string_msg macros to force explicit semi-colons in the code (or compilation failure if they're missing). This is intended to help to remedy cases in which the beautifier fails due to missing semi-colons. For developers, this means that you need to have a semi-colon after a runtime_assert, runtime_assert_msg, or runtime_assert_string_msg, just as you would if you were making a function call. commit f386b3c664c2f04af07671c8b7285242810966c0 Date: Mon Aug 24 11:24:56 2015 -0700 Fix build for crystal refinement commit ab934328fd8be0a54214395a33d1d6d1aba83ce1 Merge: 9d313db dd586c9 Date: Mon Aug 24 11:15:38 2015 -0600 Merge pull request #732 from RosettaCommons/sergey/f Adding specialized operator<< with visibility check for Tracer and TracerProxy (Thanks to @roccomoretti for the idea!) commit 9d313db628909b93422937001e171d49e892bd91 Date: Mon Aug 24 09:59:22 2015 -0500 Fix a few new cppcheck issues. commit 47f3eb3ae99758813482370e1b343ef07c21ad5a Merge: ce7b1d4 8e68e2d Date: Sun Aug 23 01:18:58 2015 -0700 Merge pull request #700 from RosettaCommons/vmullig/gellman_helix The main purpose of this merge is to fix packing when mainchain chiral centres switch their chirality (e.g. design with both D- and L-amino acids). This also introduces a subset of the L-gamma-4 amino acids with canonical side-chains. The following integration test changes are expected: - flexpepdock (There really was an HA atom in a wonky place in this integration test, which this change fixes.) - dcys_disulfidize (Expected changes to HA placement when designing with D and L residues -- the major bugfix that this merge introduces.) - cyclization (Very small changes to post-relaxation geometry with peptoid residues. Tracked this down to the swapping of 1HA and 2HA atoms that were in the incorrect relative arrangement. So that's good.) - oop_design (No changes to structure, but minor changes to some output information. Also not surprising.) - d_workflow (HA atoms are now in the correct positions after design. Yay!) - cluster_alns (Very minor changes to side-chain rotamers. Curious.) - oop_dock_design (No structural changes; minor changes to output information.) - ncaa_fixbb (Different non-canonical side-chain no longer chosen after fixing HLU. The only difference is that 1HA2 and 2HA2 in the "601" residue, which were staring out in swapped positions, are now correctly swapped back by the packer. So that's good.) commit ce7b1d45d96946ba4f9ad91630c7476045a99a4c Merge: af9e41e 7d5388b Date: Sat Aug 22 12:36:20 2015 -0500 Merge pull request #729 from RosettaCommons/roccomoretti/cppcheck_output Cppcheck output fix This is accomplished by splitting the stdout and stderr streams in parrallel.py, and ignoring the stdout one from the cppcheck runs. commit af9e41ec0fe60a2d4ab480e1bf21fdcd224ceddf Date: Fri Aug 21 23:54:53 2015 -0700 Fixing library levels in Frank's recent commit. commit b494ffe9b8881bfb4fbf3ba37fca1b1058b09300 Date: Fri Aug 21 23:05:50 2015 -0700 Fixing the build -- Frank left an unused variable in LK_BallEnergy.cc. commit 36769af988c2db6397e1e5b9b5539fe669782f84 Date: Fri Aug 21 19:56:18 2015 -0700 LKball now works with nblist_auto_update commit 57d4ed73e156cba74f294db0655d033d7f33a1b7 Date: Fri Aug 21 10:40:25 2015 -0400 Beautifying master. This is a large commit and will undoubtedly produce conflicts with everyone seeking to merge into master. The merge process for advancing past this commit is as follows: > git pull && git checkout 3f02864 > git merge 6d44fcf00 dismiss them cd into Rosetta/main/source run the python beautifier script: > python ../../tools/python_cc_reader/beautify_rosetta.py -j 40 --overwrite (or leave off the --overwrite flag to write the beautified X.cc file to X.cc.beaut, if you'd first like to see what the beautifier does) The -j flag tells the script how many processors to use. It takes about 5 minutes. SHA-1 in this commit log message because this commit log message goes into determining the SHA-1!) using the -Xours flag: > git merge -Xours and must be carefully handled. commit 6d44fcf00c51b4f525fa3e54ac581ff2d7344674 Merge: dccc86a 641b2f9 Date: Fri Aug 21 08:34:03 2015 +0300 Merge pull request #727 from RosettaCommons/sjf/parse_resnum_in_PreventResiduesFromRepackingTaskop PreventResiduesFromRepacking task operation will now allow specifying… commit dccc86acb441c47b813ee86a2bb9a791212403b9 Merge: 5cbc6dc 01965c2 Date: Thu Aug 20 21:08:39 2015 -0700 Merge pull request #724 from RosettaCommons/vmullig/gaurav_dcys Adding support for D-cysteine disulfides to the Disulfidize mover. In the process, I caught a major bug in the DisulfideMatchingPotential that was greatly limiting its usefulness for finding new disulfides (e.g. in Remodel or in Disulfidize). I recommend that users who use these protocols to place disulfides update to the latest master after this merge. Tasks: - Add interface items to mover. - Alter mover's process_pose() function. - Alter find_possible_disulfides() function. - Alter construct_poly_ala_pose(). - Alter check_disulfide_score and build_and_score_disulfide(). - Alter core::scoring::disulfides::DisulfideMatchingPotential::score_disulfide(). - Track down and fix weird DCS "atom doesn't exist" bug. - Integration test. - Update documentation. - Tweak unit test appropriately, since it now finds far more disulfides and could actually run out of time/RAM. Changes to integration tests expected: - dcys_disulfidize (Just added; expected to fail.) - rs_loophash (Failing to run on test server.) - remodel_disulfides (Previously found no disulfides; with the fix to DisulfideMatchPotential, it now finds a decent-looking one, so that's good.) - hotspot_graft (Was failing to run on test server.) - minimize_with_electron_density (As far as I can tell, this is related to the preservation of Residue information since cysteines are not being rebuilt unnecessarily any more. Very subtle numeric differences.) - place_simultaneously (Was failing to run on the test server.) - make_symdef_file (Identical PDB output, but the CYS residues, which had 0 in their temperature column before, now have values. All other residues had temperature factors that were nonzero before and after this merge. This is an unanticipated but now-expected benefit of the changes to change_cys_state(), which no longer replaces CYS residues with new Residue objects (lacking the attached data) unless it's actually switching CYS to CYD or CYD to CYS.) - database_jd2_io (Like make_symdef_file -- identical output, but values in the temperature column for CYS.) - repack_with_electron_density (Same thing again -- temperature factors for CYS.) - resource_database_locator (Same thing again -- temperature factors for CYS.) - symmetry_data_resource (Same thing again -- temperature factors for CYS.) - resource_database_tables (Same thing again -- temperature factors for CYS.) commit 5cbc6dc65a2b90781c9012ad75a7f23004eea723 Date: Thu Aug 20 16:03:14 2015 -0400 Further increasing failure threshold for performance tests commit 33998c13f1a7e19a648c0565638655bcb1b4f71a Merge: e0bcbca 2f6f186 Date: Thu Aug 20 11:50:22 2015 -0400 Merge pull request #714 from RosettaCommons/aleaverfay/prebeauty_cleanup Merging branch aleaverfay/prebeauty_cleanup into master (pull request #714). Modifying files ahead of beautification that contain un-scoped if blocks and other statements all on a single line. The beautifier now flags such lines as sources of possible bugs. However, it demands that if you want to put several statements on one line, that you should scope the if block. This commit goes through all the cases in the code that the beautifier flags and scopes them appropriately. commit e0bcbca9458ae3a265271a26de368ac4159f6e85 Merge: 602a0ba aedebca Date: Thu Aug 20 09:40:21 2015 -0500 Merge pull request #720 from RosettaCommons/roccomoretti/stored_rotamers Add the ability to load rotamers from the SDF files passed to -extra_res_mol By default, all the entries in an SDF file with the same molecule name (the first line) will be loaded as rotamers, using (indirectly) SingleLigandRotamerLibrary. To facilitate usage, if the name line is blank, it's assumed that the entry is a rotamer for the previous entry. Note that like molfile_to_params.py, only the first entry of a given name will control the ResidueType layout - all the additional entries are used for coordinates *only*. Behavior is undefined if the atom indexes of the multiple entries used for rotamers don't match up. I think *cross fingers* that this is the major outstanding issue for using SDF files directly, rather than passing them through (standard) molfile_to_params.py. (mol2 files, polymeric residues, and controlling non-default behavior are still not implemented.) commit 602a0ba3651c08d3a4f2d4bef6f69fa76f4810a7 Merge: 7430c11 8d6728d Date: Wed Aug 19 21:29:01 2015 -0700 Merge pull request #725 from RosettaCommons/rhiju/remove_override_rsd_type_limit removing rsd_type_limit warning to encourage folks to check in patches… All changed tests are the expected ones (the ones which displayed the rsd_type warning), ones that are broken (cppcheck), or ones that have been unstable recently (match & cst_to_theozyme in mac but not linux integration tests). commit 7430c11d6d7b5e373d5c842c00c90b135a8e906a Merge: 7dcad25 c379a78 Date: Wed Aug 19 21:11:24 2015 -0700 Merge pull request #723 from RosettaCommons/rhiju/remove_backtrace_in_pose_from_annotated_sequence removing print_backtrace from disulfide backwards compatibility code … only changes are in expected test (rs_loophash) and broken tests (hotspot_graft, place_simultaneously). Some match tests are giving change in output order on mac.clang.integration (but not linux) -- this appears to date to my on_th_fly_residue_types merge, and I am looking into it separately. commit 7dcad257b0c5beaa2445fd17e7f9bf41d57b2480 Merge: 8b106f1 06d9df1 Date: Wed Aug 19 22:23:56 2015 -0400 Merge pull request #715 from RosettaCommons/everyday847/who_knows_torsion_stuff A couple of fun demos of noncanonical functionality. Integration test changes are all prior failures or cosmetic changes (indeed, changes to tracer output I had added in the first place, or a bit of TR.Trace output). New vancomycin integration test, too! commit 8b106f1dc3a00c4a98dfdf2977d5c387e77d1f89 Date: Wed Aug 19 19:12:33 2015 -0400 Updating performance benchmark: - Decreasing individual tests runtime to 64 - Increasing failure tolerance to 16% commit fad00d08017dc48170012604e166a850d32513c6 Merge: 5211cc3 3907033 Date: Wed Aug 19 14:18:28 2015 -0400 Merge pull request #728 from RosettaCommons/aleaverfay/fix_coordcst_clone_w_func Adding a clone( FuncOP ) method for the CoordinateConstraint class to fix the hotspot_graft and place_simultaneously integration tests. commit 5211cc38c81a1eebff8bc5e3ac25bc8cce7c1188 Merge: 9a078ea 434f6a2 Date: Wed Aug 19 11:02:22 2015 -0400 Merge pull request #713 from RosettaCommons/JWLabonte/sugars/loops2 CCD Loop Closure/Carbohydrates: making CCD work with polysaccharides One of the main reasons I helped @weitzner refactor CCD last year (#51) was to set up a better framework that would allow one to close saccharide loops as well. However, there was still a collection of hard-coded assumptions deep in the code that prevented sugar loops from working with cut-point variants. This merge seeks to remove those assumptions, freeing the new CCD to do its job on carbohydrate loops, such as for the project I briefly described at the end of my talk at RosettaCON. This merge also fixes some silent bugs regarding ring conformations. All unit tests pass. CCD and carbohydrates integration test changes expected. code_duplication test is from new whitespace. cppcheck was not functioning properly at the time of this merge. hotspot_graph, place_siumltaneously, inverse_rotamer_remodel, and rs_loophash are all still broken. commit 9a078ea78af266d641a2bc6f573627b92703c942 Date: Tue Aug 18 21:26:14 2015 -0700 Fixing Andy's pilot app to fix the build. commit e75b6261b7a511f98673e0497f8468e7eff8d397 Merge: ad47d81 73f0a15 Date: Tue Aug 18 19:41:43 2015 -0700 Merge pull request #730 from RosettaCommons/vmullig/bettercluster_memory Fixed bug in bettercluster.cc Fixed a segfault in a pilot app. No test changes expected (or possible, really). Skipping tests. commit ad47d810f071272e89eeea8e729b8c815f8b7e4d Date: Tue Aug 18 21:49:16 2015 -0400 Fixing app_exception_handling my bad! commit 1d19e8db2020b5d0fb586c2762d282beb4e47897 Date: Tue Aug 18 14:09:00 2015 -0700 changing z1 from 1.570660756980 to 1.51 (for backward-compatibility) in the BundleGridSampler param file. No test change expected commit 225fafa89f3edcf4e7401bdc5cb22e3a9b5b2086 Merge: 9ae52c6 49c6a15 Date: Tue Aug 18 13:12:56 2015 -0400 Merge pull request #674 from RosettaCommons/everyday847/literally_cant_even So this went a bit off-track, because the assumption in question is kind of hard to track and relies on a really good understanding of what could go wrong with pseudobonds, which I do not have. Instead, I reworked some really bad (kind of related) assumptions with the HBS patches which were causing clashes and weird geometry and other bad stuff--I think we can safely use HBS dock-design with talaris now, as long as we aren't using any NCAAs. I also fixed the hbs and a3b_hbs creators so they work... much better. commit 9ae52c6b2187775c8685a250f7e09f007d211c92 Date: Tue Aug 18 11:55:06 2015 -0400 Updating parallel.py script so it communicate with child process even if no output file was set commit 7493a73399aac953e99cd9fd55989bd6eb787c60 Date: Mon Aug 17 19:33:01 2015 -0500 Fix benchmark/util/parallel.py issue. parallel.py wasn't properly waiting for subprocesses to finish, leading to a process explosion. commit 537b470630dcc983d4bd0adc943ec578862b63ca Merge: 63cc95a c497206 Date: Mon Aug 17 14:23:45 2015 -0700 Merge pull request #711 from RosettaCommons/vmullig/fix_hotspot_graft Fixing syntax in HbondsToResidueFilter in integration tests Fixing problem with HbondsToResidueFilter in hotspot_graft and place_simultaneously integration tests due to changed syntax in this filter. Note that this does NOT fix these tests -- there was something else broken affecting them. I'll merge this fix since Rebecca closed her branch fixing these tests. We might as well solve this problem so that we know when the other problem is solved. commit 63cc95ac7277ad06437162749bfa97a6a5b0b3e0 Merge: f642081 0dec207 Date: Mon Aug 17 11:12:25 2015 -0700 Merge pull request #688 from RosettaCommons/jadolfbr/db_with_path Allow pathing to database output This merge allows -out:path:all to be used by -out:use_database and the FeaturesReporter framework. Adds another option -out:path:db to be used specifically for databases. Tested using pdb output to db via score_jd2 and a feature reporter script. Fixes bug/feature request 359: https://carbon.structbio.vanderbilt.edu/mantisbt/view.php?id=359 commit f6420813df6b63e20e8b88da9fa321735c65197b Merge: 1a95fa6 567a66f Date: Mon Aug 17 09:23:24 2015 -0700 Merge pull request #719 from RosettaCommons/rhiju/fix_carbohydrate_patches_with_on_the_fly_residue_types Rhiju/fix carbohydrate patches with on the fly residue types carbohydrates integration test restored, confirmed by @JWLabonte. All other tests pass except ones that failed before (cppcheck, 3 integration tests; the match tests, which can give different output orders). commit 1a95fa68dd628f7fcf6d90996c1f9263a8c9e4e9 Merge: ce8c554 dda7198 Date: Mon Aug 17 10:45:39 2015 -0400 Merge pull request #710 from RosettaCommons/rfalford12/cleanup_meminfo_and_addmemb Updating Documentation & Organization of AddMembraneMover and MembraneInfo Its been about a year since I've updated the in-code documentation (@brief, @detials, class description) for AddMembraneMover and MembraneInfo but a lot has changed. Its also becoming harder to figure out where to add things in terms of (1) where data gets stored and (2) at what step to initialize new data in AddMembraneMover. In this pull request, I updated the documentation for both classes. I also reorganized methods so adding new extensions to these classes is easier. I've also updated some logic in AddMembraneMover to fail if none of the current edge cases cover an input PDB which is more informative to the user Integration test Changes: Code duplication: Catches consistent initialization between constructors (this is good duplication!) and mp_relax_w_ligand is cosmetic commit ce8c5546316b46e1d3a2c47bf04038fc0d06a814 Merge: ffc4590 7e27ec9 Date: Sun Aug 16 22:51:24 2015 -0400 Merge pull request #675 from RosettaCommons/everyday847/nu_angles_for_ncaas It isn't PERFECT, but it is a necessary step to get where we're going. A later (and longer) effort: synergize ring conformer and rotamer enumeration so the two are sampled together. commit ffc4590a17f54dfab5ec271b52ab7d9946dc1350 Merge: b6318de 0e69774 Date: Sat Aug 15 15:10:28 2015 -0700 Merge pull request #591 from RosettaCommons/rhiju/on_the_fly_residue_types Rhiju/on the fly residue types Link to preRosettaCon2015 talk on this pull request, including benchmarks showing significantly reduced load time: https://dl.dropboxusercontent.com/u/21569020/Das_preRosettaCon2015_ResidueTypeSet.pdf Yes finally all tests pass. Notes on changed integration tests: Cosmetic: fiber_diffraction_fad, fold_and_dock, rs_loophash, Enzrevert_xml, enzdes, mp_find_interface, match_1n9l Previously broken: hotspot_graft, inverse_rotamer_remodel, place_simultaneously (also mac.pyrosetta was previously broken) commit b6318dea132d73b2091651d6916f3f532e6a1530 Merge: 5792231 23dd930 Date: Fri Aug 14 12:56:54 2015 -0700 Merge pull request #707 from RosettaCommons/tlinsky/tomponents A few cleanups and stability fixes to StructureData commit 5792231c81b5cc21a4098cf90b7e50a2829363af Merge: ea8e574 dfcdbff Date: Fri Aug 14 14:49:55 2015 -0400 Merge pull request #709 from RosettaCommons/JWLabonte/sugars/loops Patch System: adding new PatchOperation DELETE_VARIANT_TYPE This merge allows one to remove a VariantType from a ResidueType, expanding all the fun things we can do with the patch system. I also needed it to make modified sialic acids, which are key to eukaryotic glycobiology. also added a new integration test to the carbohydrates tests to demonstrate carbohydrates test changes expected commit ea8e574bac2fcc0d324576cf9f45b1c8afb986de Date: Fri Aug 14 13:05:28 2015 -0400 Disabling T800_ClassicRelax PyRosetta unit test on Mac commit 2226e40e287a8c9907d1503da4e98b300f500de5 Merge: cf52029 dd1ecb1 Date: Fri Aug 14 11:47:56 2015 -0400 Merge pull request #694 from RosettaCommons/JWLabonte/sugars/database Carbohydrates: adding more sugars to database This merge simply adds some more sugars and a fancy bacterial modification patch to the database. In the process, I fixed tracers in `PatchOperation.cc`, which affects output in two integration tests that have the tracer volume jacked up. (Why they do, I do not know.) commit cf52029ff29a797fb4338df65d687edfbf3e823a Date: Thu Aug 13 19:32:17 2015 -0400 Updating PyRosetta release script, reducing number-of-revisions-to-keep to 1. commit 874560b36d1ea7fbff5afd3e2d486e8db7beb77c Merge: 6cf5de2 6b2f372 Date: Thu Aug 13 19:28:56 2015 -0400 PyRosetta: Disabling unit tests G000_Main_GUI, T400_Refinement and Workshop4test on Mac platform due to memory failure. commit 6cf5de2160335e0cc43e4acec76fcbefed2e313c Merge: 3eb05ff 917c95a Date: Thu Aug 13 14:33:17 2015 -0700 Merge pull request #704 from RosettaCommons/rhiju/fix_patch_selectors_nucleobase_sample_around turn on patch selectors that are needed in nucleobase_sample_around. … The problem tests are the three 'usual' broken integration tests & mac.pyrosetta (also now broken). commit 3eb05ff06277c112f840612ba508a5c95fa71664 Merge: 3b059b6 0f81130 Date: Thu Aug 13 09:57:19 2015 -0600 Merge pull request #705 from RosettaCommons/benchmark Adding parallel_execute helper function. Updating scientific test template. commit 3b059b6f2b9bf66ba7a334b32c6e8bfbb09a0d52 Merge: 0e84f3d 0ef272c Date: Thu Aug 13 09:56:25 2015 -0600 Merge pull request #706 from RosettaCommons/sergey/f Adjusting CMake config: adding CXX11 build, making default Clang build on Mac to use -std=c++11 to match our scons settings. commit 0e84f3df68b00052301d80d4e9e58042d48a249d Merge: 832c53e 21db94f Date: Thu Aug 13 09:06:04 2015 -0400 Merge pull request #701 from RosettaCommons/weitzner/dock_partners_flag Dock partners in any order ==================== This change set enables docking chains to be specified in any order. The previous functionality required that the dock_partners flag listed the chains in the same order as in the PDB file. For example, if you had a PDB file with the chains A, B and C, you could dock the A–B dimer to C ("AB_C") or A to the B–C dimer ("A_BC"), but not the A–C dimer to B ("AC_B"). To use "AC_B", the user would have to rearrange the chains in the PDB file. Absurd! :angry: This change set fixes that using `ChainSelectors` and representing dock partners as `vector1< bool >`s. There were many flavors of the `setup_foldtree` function, several of which were not called anywhere in the code. These functions have been removed with the author's permission as appropriate. Unit tests: ------------- All unit tests pass. Unit tests that tested functions that were removed were also removed and the `DockingProtocolFunctions.u` file was updated to account for `FoldTree` edges being printed in a different order. A new unit test has been added to demonstrate and test the out-of-order functionality. Integration test changes: --------------------------------- Seventeen integration tests differ (3 were failing on the branched revision, 2 are failing as a result of an unknown bug somewhere else) as a result of this change, but the differences are limited to the order in which `FoldTree` `Edge`s are printed out and do not affect the folding direction of the pose or cause trajectory changes. ### Tests with text-based differences ### These differences are all of the form: < protocols.docking.DockingProtocol: FOLD_TREE EDGE 1 71 -1 EDGE 71 107 -1 EDGE 71 126 1 EDGE 126 193 -1 EDGE 126 108 -1 --- > protocols.docking.DockingProtocol: FOLD_TREE EDGE 1 71 -1 EDGE 71 107 -1 EDGE 71 126 1 EDGE 126 108 -1 EDGE 126 193 -1 1. `docking_local_refine_min` 2. `docking_low_res` 3. `score_only_silence` 4. `fuzzy` 5. `assemble_domains_jd2` 6. `docking_local_refine` 7. `rotamer_probability` 8. `docking_distance_constraints` 9. `docking_site_constraints` 10. `docking_full_protocol` 11. `score12_docking` 12. `docking_ensemble` ### Tests that were failing on the merge base with master: 1. `place_simultaneously` 2. `hotspot_graft` 3. `inverse_rotamer_remodel` ### New test failures: ### 1. `mp_dock_prepack` (does not run) – ***disabled with permission from JKL*** 2. `mp_find_interface` (does not run) – ***disabled with permission from JKL*** commit 832c53e5fd14df0aa103997a921b2935cff5dae4 Merge: 418b1df af8600d Date: Wed Aug 12 19:16:18 2015 -0700 Merge pull request #696 from RosettaCommons/tlinsky/ft_slide_jump Fix slide_jump() bug in fold tree where root could be changed commit 418b1dfeb690162cb32f3733c0aa4afb86605b06 Merge: 380a9b2 e341ad7 Date: Wed Aug 12 21:50:03 2015 -0400 Merge pull request #702 from RosettaCommons/JWLabonte/sugars/input Carbohydrates: making Rosetta correct PDB files created by GLYCAM on input GLYCAM is a popular MD tool for carbohydrates. Unfortunately, it has its own three-letter code and treats the HO- group or RO- group of reducing end residues as their own residues. This merge will try and have Rosetta fix this problem automatically with the -glycam_pdb_format flag. This merge will also make it so that one no longer has to use the -override_rsd_type_limit with sugars. -include_sugars will set that option automatically. only expected carbohydrates test changes from changed flag and input files commit 380a9b2f9f0cacf221da5e803062724d704ce2ee Merge: 89fa80c dd0fa78 Date: Wed Aug 12 10:55:24 2015 -0400 Merge pull request #648 from RosettaCommons/jkleman/mp_updates Jkleman/mp updates SASA: * added functions in core/scoring/sasa/util that (1) compute the total per-residue-SASA, (2) relative per-residue-SASA, and (3) is_residue_exposed boolean function; this is all based on Jared's code * added sasa calculator pilot app that uses those functions: pilot/jkleman/per_residue_sc_sasa.cc * added integration test Apps: * add helix_from_sequence app with integration test, works both for soluble as well as membrane helices * cleaned up interface_statistics app for membrane proteins, still more to refine Docking: * MPDockingSetupMover: added options to optimize membrane positions for both partners before concatenation * MPFindInterfaceMover: cleaned up, added options, checking in the code from before Gordon conference, still needs work * MPQuickRelaxMover: cleaned up, added options, checking in code from before Gordon conference, still needs work RosettaMP: * MembraneInfo: added function in_membrane that returns boolean * SpanningTopology: fixed bug in concatenating topology objects that was not caught by unit test, included that in unit test now * Embedding: added functions for add_span_embedding * TransformIntoMembraneMover: added option to optimize embedding, still needs test * TranslationRotationMover: removed unused fullatom variable and option that was never used * VisualizeEmbeddingMover: debugged, but more debugging to do * most membrane movers: added check at the beginning of the apply function to check the foldtree, after apply the foldtree is reset to the original one * fixed a whole bunch of bugs, unit test oversights and integration test changes that weren't caught in the renaming of the integration tests for the membrane framework * cleaned up some of the center/normal tags and commandline initializations that didn't yet use Rebecca's function * protocols/membrane/util: added a whole bunch of useful functions: create_membrane_foldtree with anchor points at chain COMs, create_membrane_foldtree with anchor points at COMs of TM regions of chains, helper function for both, chain COM, chain TM COM, rsd closest to chain COM, rsd closest to chain TM COM, split topology by chain, compute embeddings by chain; added unit tests for all of those Other: * pose/util: added get_chains function Unit test failures: none Integration test changes: mostly cosmetic, helix_from_sequence and per_res_sc_sasa didn't exist, mp_mutate_relax and mp_quickrelax changed because the protocol changed commit 89fa80ca27a5ed4b53df0e00faa1f52a854ac3c9 Merge: 7e664da dbc5b20 Date: Tue Aug 11 18:22:51 2015 -0700 Merge pull request #703 from RosettaCommons/vmullig/pull_request_demo Updating a comment in a pilot app. This is to demonstrate how to make a pull request. You can make lists: Like this. commit 7e664da85f5ce30616af9d17793e8079956b80fc Merge: ce2f00c 947aeae Date: Tue Aug 11 10:27:22 2015 -0500 Update cppcheck scripts. * Better error reporting if scripts encounter errors * Allow for "ignoring" of historical issues. * Better debugging output commit ce2f00cea247c0e3d6fab9f1f27d9b93b13f6753 Merge: 650e766 95321d8 Date: Sun Aug 9 22:59:16 2015 -0400 Merge pull request #695 from RosettaCommons/rfalford12/remove_old_rm_code Removing old resource manager hooks in the membrane framework ---------------------------------------------------------------------------------------------- When the membrane framework was in development, we added hooks to the resource manager. However, development stopped on these. Currently, the code does not compile, and if it did, it would not work with the current version of the RosettaMP framework. Once we update to JD3/resource manager, I'll try this again but probably from scratch. No reason for non-compiling, and non-functional files to complicate master. commit 650e766d8d5c184c3fd554a78d039629cbc404a1 Merge: a8ad342 ce5ec51 Date: Fri Aug 7 16:23:09 2015 -0700 Merge pull request #691 from RosettaCommons/vmullig/gellman_helix Fixed a minor bug with helical bundle generation Fixed a minor bug with helical bundle generation when the repeating unit is more than one residue. The number of mainchain atoms in the last (overhang) residue was being checked incorrectly. Also, adding a feature to the fit_helixparams app that lets the user specify that certain bond angles or bond lengths are non-ideal. ...Aaaaand in the process I caught two more bugs in the BundleGridSampler. commit a8ad342e141a2dae244df8a7fc46f1505ccaeaaf Merge: 657466f ca43944 Date: Fri Aug 7 11:05:08 2015 -0700 Merge pull request #679 from RosettaCommons/tlinsky/tomponents API improvements to denovo_design base classes commit 657466faf43bfb2895c8d67bb75e3f9b8fcf3450 Merge: 0105408 ec05681 Date: Fri Aug 7 11:00:36 2015 -0700 Merge pull request #686 from RosettaCommons/jadolfbr/antibody_design Jadolfbr/antibody design A few small changes that should be merged. Antibody Related ---------------------- 1) Disable cartesian min during graft adaptation in the designer. To close 100% of grafts, it is not needed and can cause downstream effects from the inaccurate G. From now on, this is off to increase stability of large cluster runs. 2) Change CDR length max to 25, which is reasonable. Really, this only effects H3. 3) Add optional output of monte carlo acceptance rates to PDB file during antibody design runs. 4) Change docking during design from LH_A to A_LH so that LH is moving relative to A. Non-Antibody Related --------------------------- 1) Remove the text of what the '-correct' option should do. Its all BS. Even the defaults in the options file can be overruled through code. At least this is more clear and points people to score_function_corrections.cc. Test Changes ------------------ Expected in antibody design mover due to the change in defaults. commit 010540855af75accca97020229704a1e048fbcc1 Merge: 2e5d1c1 fcba717 Date: Fri Aug 7 10:49:50 2015 -0700 Merge pull request #684 from RosettaCommons/jadolfbr/proto_cdr4 DE Loop/Proto CDR4 analysis for the Antibody Features Reporters ----------------------------------------------------------------------------------------- Much of this work, especially in terms of the Features R framework was done by Simon Kelow (@skelow). This merge enables the Antibody Features Reporter and associated R scripts to analyze the DE loop of antibodies, what we are also calling proto_cdr4. Both names are not ideal, but for now, this is the way they are. A few functions in AntibodyInfo were refactored in order to return information on this loop as if it was a CDR. This will enable the full antibody framework to work with this loop, and has allowed us to use the implemented feature reporters without too many changes. A set of antibody-specific options have been added to the features R framework to enable global use of these loops during antibody analysis. Documentation on the wiki will be updated as this merge happens. No integration or unit test changes seen except the currently failing ones. commit 2e5d1c155a3a328a9ed032bc928b0847e9f896fb Merge: 4b1ffe6 8cfd620 Date: Fri Aug 7 10:43:39 2015 -0700 Merge pull request #682 from RosettaCommons/jadolfbr/ab_rs_debug Merge jadolfbr/ab_rs_debug This fixes Rosetta Script integration for the AntibodyDesignMover and the antibody TaskOps. commit 4b1ffe616bd4a7cd835124d3a8186f5fa18c11b6 Merge: ec57541 3be822f Date: Fri Aug 7 08:52:51 2015 -0600 Merge pull request #689 from RosettaCommons/PyRosetta Refactoring PyRosetta monolith linking so it use shorter command line. This should fix recent PyRosetta build errors on Mac. commit ec5754152c0509e330005820173df47cd9fb2e7c Merge: 5cca4e1 bf9dd25 Date: Fri Aug 7 08:24:38 2015 +0300 Merge pull request #685 from RosettaCommons/sjf/symm_filterscan_fix Sjf/symm filterscan fix commit 5cca4e1e24e895342b3304427696c5f05ad518bf Merge: 075aaa1 869a8a9 Date: Thu Aug 6 14:01:40 2015 -0700 Merge pull request #690 from RosettaCommons/vmullig/d_workflow Adding integration test for D-amino acid workflow. Tests LayerDesign, the BundleGridSampler, MutateResidue, resfiles, and PDB output. commit 075aaa1e1f1413cb6c19171a2bbec335537e4d8b Merge: 34a9ced 0fea165 Date: Thu Aug 6 11:22:48 2015 -0400 Merge pull request #681 from RosettaCommons/rfalford12/remove_membrane_types Remove core/conformation/membrane/types.hh Part II of pull requests that remove hardcoded values from the membrane framework Previously, the types.hh file included static variables that were placed throughout the code. However, this discouraged use of MembraneInfo and lead to inconsistency. New convention: The defaults should be initialized in AddMembraneMover and MembraneInfo only If you need & expect these values, grab them from the pose if you can Otherwise, the values you are created are likely not related to the pose so initialize them from scratch (rare case) I also ended up removing a huge code block from MembranePotential during my parsing.... because impulse to keep the code clean :) Note - this pull requests depends on #677!!!! commit 34a9ced5b3953b4c2e0d182ecd6f31b32c44feae Merge: 2062f13 df0cac7 Date: Thu Aug 6 02:30:22 2015 -0400 Merge pull request #687 from RosettaCommons/JWLabonte/scripts_directory_move Consolidating directories where one can place Pythons scripts per discussion at PreRosettaCON 2015... no test changes commit 2062f13a4477b9c4533398f61da06a8dd6eb71aa Date: Wed Aug 5 19:26:18 2015 -0400 Fixing performance tests comparison so it does not accumulate old info. Adding ‘profile’ benchmark script. Increasing performance tests runtime to improve stability. commit c5f849e3741fcf69a0599322b8d8ce09cea42f87 Merge: 4922e53 54452ad Date: Wed Aug 5 19:26:06 2015 -0400 Merge branch 'master' of https://github.com/RosettaCommons/main commit 4922e5356900b829812dc117cd340b54808bda45 Merge: f8e9eed ed30ea7 Date: Wed Aug 5 18:15:06 2015 -0400 Merge remote-tracking branch 'origin/benchmark' commit f8e9eedf161e04371c8cf13586aaa5e3fabe85ea Merge: 584bd5f aefc7b1 Date: Wed Aug 5 08:57:12 2015 -0400 Merge pull request #677 from RosettaCommons/rfalford12/change_default_normal2 Change membrane normal to be a unit vector by default --------------------------------------------------------------------------------- Change the default membrane normal to keep unit length instead of normalizing to 15Å. This pull request accomplishes two main goals: = Decouple the concept of the membrane normal, which represents orientation of the bilayer and the thickness of the bilayer. This is important for upcoming developments in the membrane representation = Remove hardcoding of the membrane thickness throughout the code An option is added to file_data.cc for outputting data for MEM where the normal is scaled to the thickness for visualization purposes only. To use this option, pass the flag -mp:output:normalize_to_thk. This change is now tracked by the mp_loadtime integration test. This pull request preceeds another pull request that will completely remove core/conformation/membrane/types.hh to discoruage hard-coding and increase reliance on the MembraneInfo object. Anticipated Test Changes: Fingerprint Tests: No score changes, just prints the updated normal coords. Unit Tests: Commits in this pull request update unit tests to check for a normal of unit length. Integration Tests: Anticipated changes due to changed normal length. Coordinates change in mp_relax, mp_relax with ligand, mp_dock and mp_dock_setup. We have resolved this is due to an instability in the test server and not our code. mp_loadtime also changes due to the addition described above. The following integration tests are expected to change: mp_loadtime, mp_relax, mp_relax_w_ligand, mp_dock, mp_dock_setup, mp_find_interface, mp_mutate_relax, mp_quick_relax, mp_score_jd2, mp_symdock, mp_symmetry_load, mp_transform The documentation has also been updated to reflect this change here: https://www.rosettacommons.org/docs/wiki/application_documentation/membrane_proteins/RosettaMP-KeyElements-MembraneRsd commit 584bd5fbed21f60dd88c3b7e99ae779735960389 Merge: 88213f6 1b6aec9 Date: Sun Aug 2 12:24:21 2015 -0500 Merge pull request #671 from RosettaCommons/havranek/amoeba_options Adjusting patches for DNA terminii to reflect actual molecules by adding commit 88213f63ba4b63f0cbe4f1b15f5fb6466aa007c9 Merge: 5a3ef6d 35bc45e Date: Fri Jul 31 16:34:05 2015 -0500 Merge pull request #670 from RosettaCommons/havranek/amoeba_options Adding commandline options for polarizable multipole electrostatics. commit 5a3ef6da42f33dd0e2a7b0f6bf15110691688b2d Merge: f943c60 f4af595 Date: Thu Jul 30 17:18:06 2015 -0700 Merge pull request #668 from RosettaCommons/vmullig/mpitest_off Disabling the bundlegridsampler_design_nstruct_mode MPI test until Disabling the bundlegridsampler_design_nstruct_mode MPI test until I can track down the cause of the instability. Sorry to have had this on for so long -- I was kind of hoping that one of Andy's cppcheck fixes would magically fix the instability, but no dice. Note: this means that MPI integration test failures should NOT be ignored any longer. The other MPI integration test DOES work without instability. commit f943c605763a7e272c9bf0444bbfa1ed8af2cac1 Date: Thu Jul 30 15:17:09 2015 -0500 Adding parameters for generalized Kirkwood radii. Database files only. No changes expected.