cctbxbb

Download

[email protected]

April 2011

2 participants
2 discussions

Re: [cctbxbb] Best way to compare incomplete f_hkl to calculated f_hkl
by Ralf W. Grosse-Kunstleve 11 Apr '11

11 Apr '11

Hi Jan, There are two steps to think about: .map_to_asu() and .common_set() or .common_sets(). Look for these methods in cctbx/regression/tst_miller.py, then add print statements to see how they work (run the command with e.g. "P31" so that it only exercisesone space group). However, the best way to compute f_calc given an array of f_obs is via f_calc = f_obs.structure_factors_from_scatterers( xray_structure).f_calc() This way the f_calc are directly compatible with the f_obs array. Ralf >________________________________ >From: Jan Marten Simons <marten(a)xtal.rwth-aachen.de> >To: cctbxbb(a)phenix-online.org >Sent: Monday, April 11, 2011 3:00 AM >Subject: Re: [cctbxbb] Best way to compare incomplete f_hkl to calculated f_hkl > >Hi, > >I'm looking at the best (=fastest) way to compare a (possibly incomplete) list >of observed reflections f_hkl_obs (or fo2_hkl_obs) to the f_hkl_calc of >calculated structure factors or intensities from a model structure. > >in miller.py there is a function to calculate a difference between f_obs and >f_calc, but it asserts that all reflections are present in f_obs: > >def f_obs_minus_f_calc(self, f_obs_factor, f_calc): >assert f_calc.is_complex_array() >assert self.indices().all_eq(f_calc.indices()) >assert self.anomalous_flag() is f_calc.anomalous_flag() >... > >In my case it's possible that some reflections could not be observed so they >should either be skipped or be assumed to be of zero intensity. As I could not >locate the definition for .all_eq(...), yet I'm not sure if this is already >handled and if it is how it is done. > >Then there is f_obs_minus_xray_structure_f_calc in miller.py as well. Would >this be a better approach if the model structure changes quite often? What if >I just got f_calc from another source (like doing some custom sped-up partial >structure factor calculation)? > >In general: What would be your advice on this with cctbx? > > >With regards, > >Dipl. Phys. > Jan M. Simons > >Institute of Crystallography >RWTH Aachen University >_______________________________________________ >cctbxbb mailing list >cctbxbb(a)phenix-online.org >http://phenix-online.org/mailman/listinfo/cctbxbb > > >

1 0

Best way to compare incomplete f_hkl to calculated f_hkl
by Jan Marten Simons 11 Apr '11

11 Apr '11

Hi, I'm looking at the best (=fastest) way to compare a (possibly incomplete) list of observed reflections f_hkl_obs (or fo2_hkl_obs) to the f_hkl_calc of calculated structure factors or intensities from a model structure. in miller.py there is a function to calculate a difference between f_obs and f_calc, but it asserts that all reflections are present in f_obs: def f_obs_minus_f_calc(self, f_obs_factor, f_calc): assert f_calc.is_complex_array() assert self.indices().all_eq(f_calc.indices()) assert self.anomalous_flag() is f_calc.anomalous_flag() ... In my case it's possible that some reflections could not be observed so they should either be skipped or be assumed to be of zero intensity. As I could not locate the definition for .all_eq(...), yet I'm not sure if this is already handled and if it is how it is done. Then there is f_obs_minus_xray_structure_f_calc in miller.py as well. Would this be a better approach if the model structure changes quite often? What if I just got f_calc from another source (like doing some custom sped-up partial structure factor calculation)? In general: What would be your advice on this with cctbx? With regards, Dipl. Phys. Jan M. Simons Institute of Crystallography RWTH Aachen University

1 0