Merge pull request #182 from RosettaCommons/vmullig/parametric
vmullig/parametric into master
This creates classes for storing parameters for parametric backbone generation. These classes are intended as general base classes for parameter storage for many different types of parametric backbone generation protocols, and to facilitate storing parameters inside Conformation objects so that different Movers can use stored parameters. Derived classes for the specific case of helical bundles generated using the Crick parameters are also implemented with ths pull request.
- Created core::conformation::parametric::Parameters base class. This is intended to store a set of parameters for generating one (likely-contiguous) piece of a conformation (e.g. one piece of secondary structure).
- Created core::conformation::parametric::ParametersSet base class. This stores a collection of Parameters objects for generating an assemblage of discontinuous pieces of a conformation (e.g. a helical bundle, a beta-barrel, etc.).
- Modified the Conformation object to contain a vector of ParametersSet owning pointers. This change is intended to be lightweight in terms of memory use. In most cases, one isn't working with parametrically-generated backbones, so the Conformation object won't store any additional data. In cases where one is working with parametrically-generated backbones, the Conformation object will link to the ParametersSet objects, which link to the Parameters objects.
- Created protocols::helical_bundle::parameters::BundleParameters and protocols::helical_bundle::parameters::BundleParametersSet classes, derived from Parameters and ParametersSet respectively. The former is meant to store the Crick parameters for a SINGLE helix in a bundle, and the latter, to store several BundleParameters objects to define a bundle (with some additional data for a bundle).
- Defined the variables that the BundleParameters class will hold.
- Rewrote the MakeBundleHelix mover's setters and getters so that now the MakeBundleHelix mover does not directly store the Crick parameters for the helix that it is generating, but instead holds an owning pointer to a BundleParameters object.
- This merge also brings in a small tweak to the bettercluster.cc pilot app to make it work with the new owning pointers.
- Had the MakeBundleHelix mover pass the BundleParameters object into the Conformation of the Pose that it operates on, so that the Crick parameters are stored for use by subsequent Movers (e.g. to perturb the conformation, to minimize in Crick parameter space, to report the current values of the Crick parameters, etc.).
- Had the MakeBundleHelix mover properly initialize the residue list in its BundleParameters object.
- Checked BundleParameters and Parameters copy constructors.
- Checked BundleParametersSet and ParametersSet copy constructors.
- Added a unit test to make sure that Conformation objects properly pass their Parameters objects into a unified ParametersSet when they are merged, and properly update links to Residue objects.
- Had the MakeBundle mover properly store a BundleParametersSet object, just as the MakeBundleHelix mover currently stores a BundleParameters object.
- Had the BundleParametersSet object store additional information associated with bundle symmetry.
- Added helix parameters for a D-amino acid alpha helix and beta strand, and for a "neutral" beta strand that could be composed of D- or L-amino acids (with perfect 180,180 phi/psi values).
Relegated to a future pull request:
X Add delta_z and curvature parameters to BundleParameters for Zibo.
X Write a sampling Mover that permits perturbation of the Crick parameters (and corresponding alteration of the backbone conformation) given a Pose that has parametrically-generated elements.
notify author
notify list [rosetta-logs@googlegroups.com]
pull of updated master after merging JK-lab into mpull of updated master after merging JK-lab into master
New integration tests added for DARC. See previous commit from JK-lab for details
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #183 from RosettaCommons/rfalford12/deep-copy-and-sfxn-errors
Rfalford12/deep copy and sfxn errors
Merging in fixes to deep copying of membrane info in Conformation and additional safety checks to the membrane energy terms
notify author
notify list [rosetta-logs@googlegroups.com]
Additional fixes and breakages.
Fix static build.
Also add exit code tesing to more integration tests. Some of which will now report magenta failures on the test server.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #179 from RosettaCommons/roccomoretti/integration_fixes
Trying to convert magenta to green on the integration tests
A bunch of failures were from the elimination of pose.update_residue_neighbors() in RosettaScripts. I added that back on a sub-protocol level where appropriate. However, as Filters get const poses they don't necessarily get a chance to update the neighbors. So I put back in the call before the filters run. I also fixed the issue with update_residue_neighbors() on an empty pose (the reason it was removed) and added a unit test for that functionality.
**Integration tests still broken (see pull request for details):**
coupled_moves & hybridization - should be fixed in other branches.
antibody_graft - Anyone from the Gray lab (or elsewhere) know how this is supposed to work?
filter_termini_near_interface - Anyone from the Kuhlman lab (or elsewhere) know how this is supposed to work?
dock_with_hotspot_place_simultaneously & inverse_rotamer_remodel - These are working on my machine. I'm not sure why they're crashing on the test server, and only on the release mode build.
notify author
notify list [rosetta-logs@googlegroups.com]
Fixing the c++11thread build; see the stackoverflow thread here:
http://stackoverflow.com/questions/26935824/gcc-gives-an-undefined-reference-error-to-static-data-members-in-templated-cla/26936998?noredirect=1#comment42418313_26936998
notify author
notify list [rosetta-logs@googlegroups.com]
Fixing input file in hybrid intergration test that was causing uninitialized read. Adding debug checking for inputs so code will fail cleanly on these errors. Removing some older unused options from hybridize. Thanks rhiju for finding this.
notify author
notify list [rosetta-logs@googlegroups.com]