[phenixbb] autobuild

Thomas C. Terwilliger terwilliger at lanl.gov
Sun Jan 9 14:54:25 PST 2011


Hi Rebecca,

The feature you are asking for is something that is planned, but does not
yet exist.

You have a couple options for now.  The easiest we have right now is to
build loops into an existing structure (i.e., phenix.fit_loops).

A second option is to use autobuild, and give it your existing model, with
rebuild_in_place=False, and it will try to add to it.  I think this is
what you did. If your model seems to get rebuilt from scratch when you do
this...it usually means that the model does not match the current map well
enough for autobuild to use it.  So a very good thing to do would be to
examine the "overall_best_denmod_map_coeffs.mtz" map file from autobuild
against your starting mode, and against your autobuild model, to see if
this is so.  You can also use phenix.get_cc_mtz_pdb with "quick=false" to
examine residue-by-residue the map-model correlation.

A third option is to run autobuild and turn off the main model-building
section (number_of_models=0).  Then you will get autobuild to build
outside the existing model, connect chains, fit loops, and fit helices. 
However really this is exactly the same as autobuild would do on the
second cycle of model-bulding anyhow.

I hope that this is somewhat helpful!
-Tom T

>> Dear phenix community.
>>
>> We would like to 'fix' the residues of our current protein model, and use
>> Autobuild (perhaps via rebuild-in-place=false, which is what we've been
>> trying) to add missing residues to our current model . However, our
>> efforts
>> thus far have only resulted in models that get rebuilt from scratch. Is it
>> possible to just 'add' residues to a current structure using Autobuild? If
>> so, is someone could share the correct command with us, we'd be very
>> appreciative. We can't seem to find it.
>>
>> Thanks in advance,
>>
>> Rebecca
>> _______________________________________________
>> phenixbb mailing list
>> phenixbb at phenix-online.org
>> http://phenix-online.org/mailman/listinfo/phenixbb
>>




More information about the phenixbb mailing list