Merge pull request #946 from RosettaCommons/vmullig/peppredict_boinc3
Fix remaining issues with changes to BOINC so that David Kim can deploy
Tasks:
-- Make background and borders more intense.
-- Fix border draw issue prior to first object draw.
-- Ensure that non-peptide prediction protocols do not use spacefill by default.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #924 from RosettaCommons/rhiju/bps_updates
Rhiju/bps updates
+ needed for farna renovation (which will be a separate pull request), and for use of base pair steps in various RNA modeling applications in Das lab.
+ changes to stepwise integration tests appear to be due to a numerical instability induced by different compiler optimization of protocols.d_5 code. Evidence for this being a negligible instability is in .swa_rna_gagu_10_prepend_and_ccd_close (where a number z = 5.89095e-14 changes to z = 5.33868e-14).
+ unit test for -extra_res_fa failed, but is getting fixed in Andy's pull request #943.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #939 from RosettaCommons/vmullig/fix_cassie_issue
Add a unit test illustrating a current issue with params files]
This merge will create a FAILING unit test, deliberately. This is to illustrate a current bug: recent changes have broken backwards-compatibility with custom params files provided by users. It seems that the L_AA property is required for the test params file to work; without it, it cannot be used.
Currently, reading a PDB file with a custom residue specified in a params file passed with the -extra_res_fa flag doesn't work if the params file doesn't specify D_AA or L_AA in its PROPERTIES line. I'll let those who are more familiar with the relevant code fix this; for now, this unit test will fail since this bug is unfixed.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #938 from RosettaCommons/rhiju/unify_rna_pair_classes
Rhiju/unify rna pair classes
only tests that change are the new one *rna_denovo_base_pair_setup* which tests cis/trans specification of RNA base pairs. The other ones are the usual ones with instabilities, and change on mac or linux but not both.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #937 from RosettaCommons/rhiju/cleanup_core_pose_rna_base_pair_class
Rhiju/cleanup core pose rna base pair class
all tests pass except for the 'usual'
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #849 from RosettaCommons/roccomoretti/pymol_obs
Make the PyMolObserver a CacheableObserver.
On certain systems there's an issue compiling because boost::bind via the observer system is expecting a raw pointer to the PyMolObserver, and it's being given an owning pointer. This works for boost shared pointers, but isn't robust for std::shared_ptr.
The naive fix of simply giving things a raw pointer to the observer doesn't work, because AddPyMolObserver et al. doesn't keep around any reference to the newly created observer object, and if the PyRosetta-level user doesn't store the return value, the observer object gets garbage collected and subsequent observer notifications mean segfaults. To solve this, I changed the PyMolObserver into a CacheableObserver, and cache it in the pose which we're observing.
One complication is that pose.clear() will remove any cached observers, resulting in the PyMolObserver being disconnected from the pose. The standard use cases look to be fixed, but it may be an issue in some other protocols. (Most notably, the PyMolObserver won't persist through loading a new PDB into the pose, as that calls pose.clear()) Another consequence is that not every snapshot from the previous runs are replicated in the new runs - but this should be mostly cosmetic.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #926 from RosettaCommons/vmullig/improve_aacomp
Add user-requested improvements to aa_composition score term
As listed in issue #901, users have requested a number of new features to make the aa_composition score term useful for _de novo_ protein design. This pull request begins to implement some of these.
Tasks:
- [x] Unify the PENALTY_DEFINITION groups. Rather than having separate code for counting amino acid types or amino acids with certain properties, let all of these be defined together.
- Details: I won't implement full Boolean logic at this point, since it's hard to do. Instead, residues will be counted in a given group if: (any TYPE matches) OR (((all PROPERTIES match) OR (any OR_PROPERTIES) OR (no TYPEs defined and no PROPERTIES defined AND no OR_PROPERTIES defined)) AND (no NOT_PROPERTIES match) AND (no NOT_TYPE matches))
- [x] Move AACompositionEnergy class to a new namespace, in which helper classes can be created.
- [x] Update unit tests to reflect new namespace.
- [x] Modify the AACompositionPropertiesSet class to store both types and properties.
- [x] Modify the AACompositionPropertiesSet class for the Boolean logic, above.
- [x] Add unit tests for the more complicated Boolean logic.
- [x] Update documentation for this.
- [x] Add clone() functionality to ResidueSelectors.
- [x] Add an AACompositionConstraint for constraining sequences, analogous to geometric constraints.
- [x] Store a selector in the object.
- [x] Store a AACompositionEnergySetup object in the object.
- [x] Have it initialize the AACompositionEnergySetup object from an input .comp file.
- [x] Functions to hand off the AACompositionEnergySetup object to the score term at scoring time.
- [x] Add a SequenceConstraint base class for the AACompositionConstraint.
- [x] Modify the ConstraintSet container to accept SequenceConstraints.
- [x] Add a mover to add sequence constraints that takes a .comp file and an optional ResidueSelector.
- [x] Strip out extraneous stuff.
- [x] parse_my_tag() should let the user set an input comp file and a selector.
- [x] Create an AACompositionConstraint at parse time; clone it and add it to the pose at apply time.
- [x] Documentation for this.
- [x] Add a mover to clear sequence constraints.
- [x] apply() function
- [x] Documentation for this.
- [x] Set the aa_composition score method to score using one or more global .comp files (if defined -- e.g. by the user, using a flag) AND one or more scorefunction-specific .comp files (if defined -- e.g. by the user, in RosettaScripts), AND .comp files attached to constraints stored in the pose (if defined -- e.g. by the user, in RosettaScripts). These last can be filtered by ResidueSelectors to allow the user to score sub-regions of a pose.
- [x] AACompositionEnergy should store a vector of AACompositionEnergySetup objects: one from global settings and many from scorefunction setup (e.g. in RosettaScripts).
- [x] Add a setup_residuearrayannealableenergy_for_packing() virtual function (that does nothing by default but which can be implemented by derived classes) to the ResidueArrayAnnealableEnergy that's the parent class for the AACompositionEnergy.
- [x] At score time, the score term can construct a vector of AACompositionEnergySetup objects and a vector of ResidueSubset masks that it will use for scoring. These can be taken from its internally-stored vector and from the pose. Add a version of the calculate_energy() function that takes these two vectors.
- [x] Have the finalize_total_energy() function generate the vectors and call the modified calculate_energy() function directly (for scoring).
- [x] Implement a setup_residuearrayannealableenergy_for_packing() function that will create the vector of AACompositionEnergySetup objects, stored temporarily in the AACompositionEnergy object, and the vector of ResidueSubset masks.
- [x] Have the default calculate_energy() function call the modified calculate_energy() function, passing it the cached vectors (for packing).
- [x] Integration test for this (that also tests the above two movers).
- [x] Documentation for this.
- [x] Fix whitespace issue identified by Zibo.
- [x] Look at bug identified by Peilong.
- [x] Add the following properties to the PROPERTIES table and to params files:
- [x] ALIPHATIC
- [x] POSITIVE_CHARGE
- [x] NEGATIVE_CHARGE
The following tasks were grouped into pull request #928, which was merged before this pull request:
- Move residue selectors to core.4 from core.3 so that constraints can use them.
- I'm moving these out of core/pack (since they don't just control the packer) and into core/select/residue_selector/.
- Resolve compile issues.
- Resolve library levels problems.
- Resolve any unit test problems.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #934 from RosettaCommons/revert-929-dimaio/rama_prepro
Revert "Dimaio/beta_nov15"
Reverts RosettaCommons/main#929
@fdimaio : I'm temporarily reverting the beta_nov15 merge until we can figure out why it breaks so many of the unit tests on Mac/clang. I'll also revert the later commit to master that altered the FACTS score test as part of this pull request, so that when you later revert this pull request, you'll get all of that back.
Let's fix whatever is wrong with the clang unit tests in a branch, then revert this reversion to put everything back into master.
notify author
notify list [rosetta-logs@googlegroups.com]