Merge pull request #4533 from RosettaCommons/roccomoretti/parse_input_refpose
Address using the input pose as a reference pose in parse_my_tag.
This PR deals with those Movers and Filters which purposely use the input pose as a reference structure. The approach I'm taking is to add the input pose to the DataMap, but in a fashion where it's (sort-of) treated like another reference pose - in fact, those Movers/Filters which use the input pose but can't already take a reference pose have been updated such that they can take a regular (SavePoseMover) reference pose.
Information about how to update things for the new scheme are documented in the "Updating RosettaScripts" page on the documentation wiki.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #4603 from RosettaCommons/JackMaguire/tensorflow_manager_pr2
Second incremental merge of the TensorFlowManager
This PR adds the ability to evaluate multiple tensors at once, which is called a "multirun".
Main branch: #4362
First increment: #4545
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #4601 from RosettaCommons/jadolfbr/fix_quench
Fix GlycanTreeModeler quench mode to pre-virtualize residues
This reduces bogus interactions of glycan trees that have yet to be modeled.
Also:
- Skip virtualization in ConvertRealToVirt mover if the residue is already virtualized.
- Remove an extra real to virtualize call in the modeler, which did nothing but very slightly increase run-time at each layer build.
@huxz
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #4622 from RosettaCommons/vmullig/tweak_atomicdepth
Small tweaks to AtomicDepth calculation
This makes the following small tweaks to the AtomicDepth calculation:
- When attempting to mutate residues to leucine, only mutate alpha-amino acids. (Would crash otherwise.)
- When attempting to mutate residues to leucine, do not mutate sidechain-connected residues. (Would crash otherwise.)
This code is used by the `approximate_buried_unsat_penalty` scoreterm. (I sadly have to admit that, unlike my `buried_unsatisfied_penalty`, the `approximate_buried_unsat_penalty` is quantum computer-ready, since it's pairwise-decomposable.)
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #4602 from RosettaCommons/jadolfbr/sugar_bb_defaults
Change sugar_bb default and add intra_rep term if not present
- Change sugar_bb weights to .5, as this is ideal for both ref_2015 and beta. Turn on `fa_intra_rep_xover4` at 0.55 if not present when using `-include_sugars`.
- Generally recommend the beta/Rosetta-ICO ScoreFunction if intra_rep is off.
All three of these (sugar_bb .5, intra_rep, and generally -beta) improve decoy discrimination significantly when de-novo modeling sugars, or building sugars into density.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #4586 from RosettaCommons/jadolfbr/disulf
Ignore virtual residues in FullAtomDisulfideEnergy
If we have virtual residues on either side of a potential disulfide, ignore it. We already do this for REPONLY, we should do this for virtual residues as well, like other energy terms.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #4613 from RosettaCommons/vmullig/update_pymol_selection_metric
Making SelectedResiduesPyMOLMetric robust to PDB files that lack PDBInfo, and providing the option of outputting Rosetta residue numbering instead of PDB residue numbering.
The SelectedResiduesPyMOLMetric currently assumes that all poses have PDBInfo. This updates the metric to be compatible with poses that don't, and adds an option to allow a user to force Rosetta numbering as output instead of PDB numbering, if he or she so wishes.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #4608 from RosettaCommons/yingyue0414-patch-quip_namefix
Quick Fix Quivonose Name
Quivonose is 6-deoxy-Glucose instead of 6-deoxy-Galactose (which is fucose)
The parameter files were all correct but the comments were not changed when converted from fucose.
For instance in to2-alpha-D-Quip
# ->2)-alpha-D-quinovpyranosyl [->2)-alpha-D-6-deoxygalactopyranosyl]
# ->2)-alpha-D-quinovpyranosyl [->2)-alpha-D-6-deoxyglucopyranosyl]
# Fucose is a common galactose variant
# Quinovose is a common glucose variant
notify author
notify list [rosetta-logs@googlegroups.com]