Merge pull request #4073 from RosettaCommons/roccomoretti/PDB_loading_adjustments
A few fixes related to PDB diagnostic test
There were a few aramid-patching issues when loading PDB with the CCD.
Also change how the C-term heuristic for polymeric CCD component loading works. There's nowhere close to a consensus on how the OXT hydrogen is named in different residues. We should just avoid special casing them altogether.
Also make it such that a patching failure doesn't necessary bring down the whole run. (Though we should still get a "failure" in PDB_diagnostic test if we hit an issue.)
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #4064 from RosettaCommons/roccomoretti/ligand_converter
A simple restype conversion utility.
One missing feature that comes up occasionally for me is the lack of a simple way of taking an arbitrary params file (or similar) and converting it into a viewable output.
This PR adds a simple utility application which allows you to specify certain residue types from the database (including patched types), CCD or command line, and output them in PDB, sdf or params file output.
Note that currently there's no guarantee that any of the output formats are able to be round-tripped through Rosetta, or that they're usable for all but the most rudimentary purposes. But it should be enough to get a PyMol-loadable pdb or a params file skeleton of a given ResidueType.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #4065 from RosettaCommons/mlnance/add_utility_exit_to_ConstraintIO
Warn user in ConstraintIO when a residue specified by a constraint file does not exist in the Pose
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #4068 from RosettaCommons/JWLabonte/quick_fix
Bug Fix: ligand_motif_design integration test
This merge will hopefully correct the `ligand_motif_design` integration test, which I accidentally broke when adding ATP to the database. (:rat: is sorry.) The test formerly used its own ATP `.params` file, and Rosetta will not allow there to be two ATPs in a `ResidueTypeSet`.
I am simply making the test use the database version, which also involves renaming the atoms in the input `.pdb` file to their correct PDB names.
All tests pass, but there are naturally differences in the `ligand_motif_design` test.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #4062 from RosettaCommons/benchmark
Benchmark scripts update
- updating scientific/command.py so it delete Python virtual environment even if test terminate abnormally
- updating PyRosetta unit test Benchmark script and impose timeout limit for duration of unit tests
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #4050 from RosettaCommons/rfalford12/membrane-sci-benchmark3
Membrane energy function scientific benchmarks part #2: decoy discrimination
--------------
This pull request will add the decoy discrimination scientific benchmark tests for membrane energy functions. This test is part of a suite of four tests that can be used to compare implicit membrane models, and have been most recently used to demonstrate sizeable scientific improvements by franklin2019.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #3930 from RosettaCommons/smlewis/green-pdb-diag
This replaces the pass/fail system for the PDB diagnostic scientific test with a 5% failure threshhold instead of the granular "everything that passed before must continue passing". The previous version is "better" but also eternally red because something always goes wrong, and then we are trained to just ignore it. All the old reporting tools remain in place for fine grained diagnostics and finding bugs to fix: the test will now only turn red if something goes wrong in a big way (thousands of PDBs start failing). The threshold is at 5% failure.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #3956 from RosettaCommons/roccomoretti/fix_clang_analysis
Comment out unneeded code in GALigandDock LigandAligner.
Clang static analysis was flagging these variables as being unused. From a brief examination of the code, it looks like there's a bunch of (currently) unnecessary calculation. It may have been used earlier for z-score calculation, but it isn't doing anything currently, so comment it out.
notify author
notify list [rosetta-logs@googlegroups.com]