[phenixbb] question about xn joint refinement

[Pavel Afonine]

Hi Qun,

- I agree with Ralf: the version you use is really old and I would 
suggest using the latest available that you can download from this web 
page: http://www.phenix-online.org/download/

- When you run joint XN refinement, you need to explicitly specify which 
data is neutron and which is X-ray (there is no (reliable) way for 
phenix.refine to figure this out automatically). So, I would correct 
your command below as following and run again:

phenix.refine xray_data.file_name=dhfr-mtx-rt-xray.hkl 
neutron_data.file_name=newdhfr_un_change.hkl xn_easy_080326a.pdb 
input.xray_data.labels=FOBSx input.neutron_data.labels=FOBSn 
xray_data.r_free_flags.generate=True 
neutron_data.r_free_flags.generate=True refine_default.params

The fraction of your file "refine_default.params" looks good to me. The 
choice for scattering dictionary is automatically adjusted given the 
input data. I agree it is a bit confusing the way it's reported in 
parameters file, I'm making a note of this right now and I will make it 
clearer in future.

A few points to keep in mind doing joint XN refinement:
- you can automatically add H atoms as well as exchangeable H/D using 
phenix.ready_set. Example:
phenix.ready_set model.pdb

- when doing joint XN refinement, try using "optimize_wxc=true 
optimize_wxu=true". This might be essential to get a good result.

- also, I would recommend running separate X-ray and Neutron refinements 
to compare with.

Please let me know if you have any questions or problems with this; I 
will be happy to help with any issues.

Cheers,
Pavel.


On 1/1/2009 9:15 PM, crystallogrphy wrote:
> Deall all:
>
> I have a question about xray and neutron joint refinement.
>
> When I start phenix joint refinement like this, phenix gave the wrong 
> message:
> phenix.refine dhfr-mtx-rt-xray.hkl neutron_data.file_name=
> newdhfr_un_change.hkl xn_easy_080326a.pdb input.xray_data.labels=FOBSx 
> input.neutron_data.labels=FOBSn xray_data.r_free_flags.generate=True 
> neutron_data.r_free_flags.generate=True refine_default.params
>
> ===================== joint neutron/xray refinement start 
> =====================
>
> Traceback (most recent call last):
>     call_back_after_monitor_collect=call_back_after_monitor_collect)
>   File "/usr/local/phenix-1.3-rc4/phenix/phenix/refinement/driver.py", 
> line 1031, in run
>     call_back_after_monitor_collect = call_back_after_monitor_collect)
>   File 
> "/usr/local/phenix-1.3-rc4/phenix/phenix/refinement/strategies.py", 
> line 184, in refinement_machine
>    * assert not neutron_refinement
> AssertionError*
>
> I guess in my default parameters, something is wrong, particularly the 
> bold word I highlighted. But I do not know how to fix. Does any one 
> have a clue?\
> Thanks and Happy New Year!
>
> main {
>     bulk_solvent_and_scale = True
>     simulated_annealing = False
>     ordered_solvent = False
>     ncs = False
>     ias = False
>     number_of_macro_cycles = 3
>     max_number_of_iterations = 25
>     use_form_factor_weights = False
>     tan_u_iso = False
>     use_convergence_test = False
>     target = *ml mlhl ml_sad ls
>     min_number_of_test_set_reflections_for_max_likelihood_target = 50
>     max_number_of_resolution_bins = 30
>     reference_xray_structure = None
>     use_experimental_phases = None
>     compute_optimal_errors = False
>     random_seed = 2679941
>     *scattering_table = wk1995 it1992 *n_gaussian neutron*
>     use_normalized_geometry_target = True
>     target_weights_only = False
>     use_f_model_scaled = False
>     max_d_min = 0.25
>     fake_f_obs = False
>
>
>
>
>
> -- 
> Qun Wan
> Department of Pharmacology, Case Western Reserve University
> 10900 Euclid Avenue, Cleveland, OH 44106
> email: qunwan1 at gmail.com <mailto:qunwan1 at gmail.com>
> lab phone: 216-368-3337
> ------------------------------------------------------------------------
>
> _______________________________________________
> phenixbb mailing list
> phenixbb at phenix-online.org
> http://www.phenix-online.org/mailman/listinfo/phenixbb
>   
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://phenix-online.org/pipermail/phenixbb/attachments/20090102/a66f57ca/attachment-0003.htm>