[phenixbb] Autobuild at 5A resolution

Thomas C. Terwilliger terwilliger at lanl.gov
Thu Sep 17 05:42:02 PDT 2009


Hi Peter,

You could do something along these lines:

1. take the model that you want fixed (perhaps RNA + existing protein
model) as "starting.pdb"

2 then run autobuild with

phenix.autobuild  data=data.mtz seq_file=seq.dat \
 input_lig_file_list="starting.pdb" \
 helices_strands_only=True \
 refine=False

where helices_strands_only=true will find the helices (and strands...I
don't think you can turn that off right now) only, but not try regular
model-building.

Alternatively, you can try:

phenix.autobuild data=perfect.mtz model=starting.pdb insert_helices=True \
helices_strands_start=False build_outside=False fit_loops=False
use_resolve_fragments=False \
connect=False skip_combine_extend=True number_of_parallel_models=1 \
seq_file=seq.dat  refine=False rebuild_in_place=False number_of_models=0 \
keep_pdb_atoms=False

which will just try to insert helices using phenix.find_helices_strands
and density modify, skipping all other build steps.

There isn't any way to conveniently use both build_outside_model and
helices_strands_only at the moment, however.

Let me know if these don't do what you need,

All the best,
Tom T

>>  Dear Phenix users,
>>
>> I am trying to solve a big RNA/protein complex at 5A resolution. Phases
>> come
>> from MR solution with a partial model in Phaser. Thanks to  NCS the map
>> after density modification in RESOLVE is of reasonable quality for this
>> resolution and new secondary structures, out side of the model, can be
>> detected.
>> I would like to improve the maps and model with something like a poor
>> man's
>> version of AutoBuild. An Autobuild for very low resolutions that does not
>> include refinement  but iteratively carries out density modification and
>> identification of helices/strands outside of the model with
>> phenix.find_helices_strand.
>> Does this sound feasible, Are there already scripts the follow such a
>> scheme
>> ?
>>
>> Many thanks in advance for your advice
>>
>>
>> --
>> Peter
>> _______________________________________________
>> phenixbb mailing list
>> phenixbb at phenix-online.org
>> http://www.phenix-online.org/mailman/listinfo/phenixbb
>>




More information about the phenixbb mailing list