Tsjerk and Ezra, thanks a lot for your thoughts. Tsjerk I see your point that an assembly is one snapshot, and for many purposes a single state object or multiple (grouped) objects will be desired. However, loading the assembly into multiple states in the first place is not a dead end and using the commands split_states, group, alter (to rename chains) and create gives you all the options to go from the multi-state assembly to any other configuration. Chain renaming is not trivial, we should provide a convenient shortcut for that. I like the multi-state solution because it's memory efficient and preserves chain ids.
Ezra, a separate command to create an assembly from an asymmetric unit object is a good suggestion (similar to http://pymolwiki.org/index.php/BiologicalUnit/Quat). So far the implementation doesn't keep the parsed mmCIF file in memory, but that would be doable, considering that typical PDBx/mmCIF files are only few megabytes in size. Cheers, Thomas On 10 Apr 2015, at 10:25, Ezra Peisach <ezra.peis...@rcsb.org> wrote: > I appreciate the plan to support biological assemblies... Initial testing > shows that it is doing the right thing. > > One disadvantage of this plan is that you need to know a priori how many > assemblies are present prior to the fetch command. > > Given that you are loading the mmCIF file into memory - which contains the > biological assembly information, I would prefer an option to generate after > the fetch command.... > > Ezra On 10 Apr 2015, at 00:59, Tsjerk Wassenaar <tsje...@gmail.com> wrote: > Hi Thomas, > > The loading of symmetry related groups as states feels a bit odd to me and > the first thing I usually do is split_states. Coming from molecular dynamics, > I usually see states as snapshots, and an assembly is still one snapshot. I > find it most natural to have the different units as objects, but since PyMOL > has grouping capabilities, why not load them as group of objects? More > complex chain identifiers could be an option, but you'd want to be able to > select all equivalent chains with a simple statement. > > Cheers, > > Tsjerk > On Thu, Apr 9, 2015 at 6:48 PM, Thomas Holder <thomas.hol...@schrodinger.com> > wrote: > Hi all, > > I'm working on reading biological assemblies from mmCIF files. The patch > against the latest SVN code can be downloaded from > http://pymol.org/tmp/pymol-assembly-svn-r4115.patch > > This patch adds a new "assembly" setting, which is the assembly id (empty/off > by default). The mmCIF reader attempts to read that assembly into coordinate > sets to mimic the multi-model pdb1 files. If the assembly has more than one > state, then all states are activated for rendering for that object (state=0 > object level setting). Assembly transformations are ignored for multi-model > structures (NMR models etc.). > > Example: > PyMOL> set assembly, 1 > PyMOL> fetch 1rmv, type=cif > > You can put the setting in your pymolrc to have PyMOL load assemblies by > default. > > Before I push this to SVN I'd be interested in your feedback. Is this a good > and user friendly solution? Just to give you an idea about other possible > approaches (which I think are more complicated or less flexible): > > - add an "assembly" argument to "load" and "fetch" > - add new functions "load_assembly" and "fetch_assembly" > - create objects instead of states for symmetry units > - create one single-state object with new chain enumeration > > Cheers, > Thomas -- Thomas Holder PyMOL Principal Developer Schrödinger, Inc. ------------------------------------------------------------------------------ BPM Camp - Free Virtual Workshop May 6th at 10am PDT/1PM EDT Develop your own process in accordance with the BPMN 2 standard Learn Process modeling best practices with Bonita BPM through live exercises http://www.bonitasoft.com/be-part-of-it/events/bpm-camp-virtual- event?utm_ source=Sourceforge_BPM_Camp_5_6_15&utm_medium=email&utm_campaign=VA_SF _______________________________________________ PyMOL-users mailing list (PyMOL-users@lists.sourceforge.net) Info Page: https://lists.sourceforge.net/lists/listinfo/pymol-users Archives: http://www.mail-archive.com/pymol-users@lists.sourceforge.net
