Re: [cctbxbb] Getting r1_factor of miller arrays with different count of miller indices

Actually, the R2-factor (or more commonly the wR2-factor) is calculated using the squared differences in the intensities rather than the squared differences in the amplitudes as I stated in my previous email. See also section 2.7 in the SHELX manual. Cheers, Richard On 6 July 2012 12:34, Richard Gildea wrote:

Hi Pavel,

Small molecule crystallographers often refer to the R-factor as the R1-factor. The '1' just means that is calculated using the amplitudes without squaring. An R2-factor would be obtained by replacing the amplitudes by the squares of the amplitudes and taking the square root. See for example equations 7 and 8 in this paper by David Watkin ( http://dx.doi.org/10.1107/S0021889808007279). The terminology R1 factor is also used in the SHELX manual.

Cheers,

Richard

On 6 July 2012 12:23, Pavel Afonine wrote:

...

Hi,

can anyone tell me what's r1_factor and how it is different from classical text-book definition of R-factor, for example:

http://en.wikipedia.org/wiki/R-factor_(crystallography)

(or for those who doesn't trust wiki: page 147 in Blundell&Johnson (1976))

?

Are there r2_factor, r3_factor, ..., etc ?

Thanks, Pavel

On 7/6/12 12:09 PM, Richard Gildea wrote:

Just to avoid any confusion between mine and Nat's responses, there are two similarly named functions in miller.array, common_set and common_sets.

Thus the following two lines (note singular common_set):

f_obs = f_obs.common_set(f_calc) f_calc = f_calc.common_set(f_obs)

is equivalent to the one-liner (note plural common_sets):

f_obs, f_calc = f_obs.common_sets(f_calc)

Cheers,

Richard

On 6 July 2012 12:04, Richard Gildea wrote:

...

Hi Jan,

You will need to get a matching set of indices for both arrays before calculating the R1 factor. The simplest way to do this would be by calling the miller.array.common_sets() function:

e.g. where f_obs and f_calc are two arrays that are not necessarily matching sets of indices:

f_obs, f_calc = f_obs.common_sets(f_calc)

Hope that helps,

Richard

On 6 July 2012 11:59, Jan Marten Simons wrote:

...

Hi,

I'm facing a new challenge while working with cctbx:

Imagine a set of (possibly incomplete) measured intesities, or integrated intensities from xrd powder patterns (I_obs) and a fitting crystal symmetry (xtal_symm). Now if I want to check if a given structure would generate the same intensities. (--> low R1) the following code exibits the problem:

# -*- coding: utf-8 -*- from __future__ import division from cctbx import xray from cctbx import crystal from cctbx.array_family import flex from libtbx import Auto

obs_file = "Test.hkl" # in shelx hklf4 format xtal_symm = crystal.symmetry( unit_cell=(4.000, 5.4321, 7.531, 90.0, 90.0, 90.0), space_group_symbol="P mm2")

trial_structure = xray.structure( special_position_settings=crystal.special_position_settings( crystal_symmetry=xtal_symm), scatterers=flex.xray_scatterer([ xray.scatterer( label="Si", site=(0.0,0.0,0.0), u=0.2)]))

# load (integral) intensities from diffraction data from iotbx import reflection_file_reader rfl = reflection_file_reader.any_reflection_file("amplitudes="+obs_file,

ensure_read_access=True) I_obs = rfl.as_miller_arrays(crystal_symmetry=xtal_symm, force_symmetry=False, merge_equivalents=True, base_array_info=None)[0] I_obs = I_obs.discard_sigmas() f_obs = I_obs.as_amplitude_array()

f_calc = trial_structure.structure_factors(d_min=1.0).f_calc() f_calc.merge_equivalents()

R1 = f_calc.r1_factor(f_obs, scale_factor=Auto, assume_index_matching=False)

gives: " assert other.indices().size() == self.indices().size() AssertionError "

Is there some way to get the R1_factor for this kind of scenario?

Cheers and thanks in advance, Jan _______________________________________________ cctbxbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/cctbxbb

_______________________________________________ cctbxbb mailing [email protected]http://phenix-online.org/mailman/listinfo/cctbxbb

_______________________________________________ cctbxbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/cctbxbb