Port Chimera's copy/combine

[sitins07@…]

As per summary

[Tristan Croll]

Adding myself to CC - this should be really useful for ISOLDE.

[Greg Couch]

Note that various metadata should be maintained when editing. For example, if a whole chain is copied and its sequence is authoritative (ie., chain.from_seqres is True), then the sequence and the fact that it is authoritative should be copied too. For residues, there is a mmcif_chain_id attribute that should be set/cleared as appropriate. There are probably other examples.

[Eric Pettersen]

Relevant information in ticket #3045

[Tristan Croll]

Making some improvements to my implementation of this (namely making 
sure that fragments internal to a chain get inserted in the right place 
rather than simply appended to the end of the residue list). One thing 
that might be worth making happen automagically: if I insert a fragment 
into a chain break but don't connect it at either end (that is, 
effectively creating two new smaller chain breaks), then the 
missing-structure pseudobonds don't update on their own. Alternatively, 
if pdb_connect::find_missing_structure_bonds() were exposed to the 
Python layer I could just have my method clear all the existing 
missing-structure pseudobonds and call that.

On 2020-04-15 16:55, ChimeraX wrote:

[Tristan Croll]

Went ahead and added some logic to my code to update the 
missing-structure pseudobonds (for protein, at least... will extend to 
nucleic acids later). By the way, I *love* the new presets feature - a 
great way to put in quick little snippets of test code when building a 
new method! The below test seems to be working fine, with one caveat: it 
still messes up the sequence viewer a bit. Doesn't crash it, but the 
deleted residues appear as missing and their replacements appear as a 
sequence appended to the end of the chain (despite the fact that they're 
inserted in the right place using the `precedes` argument of 
`new_residue()`).

{{{
from chimerax.core.commands import open as cxopen
mlist = cxopen.open(session, '1pmx')[0].child_models()
m0 = mlist[0]
m1 = mlist[1]
m2 = mlist[2]
m3 = mlist[3]
m1.display=False
m3.display=False
session.models.close(mlist[4:])
m0.residues[10:25].delete()
m2.residues[26:43].delete()
from chimerax.isolde.atomic.building.merge import merge_fragment
merge_fragment(m0, m1.residues[10:25], anchor_n = m0.residues[9], 
anchor_c=m0.residues[10])
merge_fragment(m2, m3.residues[27:42], update_style=False)
}}}






Seems to be working nicely, with the caveat that it still messes up the 
sequence viewer a bit. Doesn't crash it

On 2020-04-20 18:34, ChimeraX wrote:

[Eric Pettersen]

If the other scenarios correctly update the missing-structure pseudobonds, it would be disappointing (and mildly useless) if the inserted-but-disconnected fragment scenario didn't. May take me a few days to get to it.

Glad you're liking the sneaky use of presets. :-)

Not shocked the sequence viewer got discombobulated, but nonetheless it shouldn't. Will open a ticket.

--Eric

[Eric Pettersen]

Replying to Tristan Croll:

Making some improvements to my implementation of this (namely making 
sure that fragments internal to a chain get inserted in the right place 
rather than simply appended to the end of the residue list). One thing 
that might be worth making happen automagically: if I insert a fragment 
into a chain break but don't connect it at either end (that is, 
effectively creating two new smaller chain breaks), then the 
missing-structure pseudobonds don't update on their own. Alternatively, 
if pdb_connect::find_missing_structure_bonds() were exposed to the 
Python layer I could just have my method clear all the existing 
missing-structure pseudobonds and call that.

On 2020-04-15 16:55, ChimeraX wrote:

Tristan, this is a port of Chimera's "copy/combine" functionality, for copying or combining entire structure models -- so not really a replacement for "Isolde merge". Though I suppose you can sort of fake it by deleting large parts of one structure before the combine, with probably mixed results.

[Eric Pettersen]

'combine' command available in tomorrow's build.