- cctbxbb - phenix-online.org

Assertion Error in twin target test
by Luc Bourhis 23 Mar '07

23 Mar '07

Hi all, I have modified the least square residual classes in cctbx/xray/ targets.h and cctbx.xray to handle refinement against F^2. In the subsequent tests, the following assertion failure cropped up. It does not feel to me my changes caused it… Luc libtbx.python /Users/luc/Developer/cctbx/cctbx_sources/cctbx/cctbx/ regression/tst_twin_target.py approx_equal eps: 0.5 approx_equal multiplier: 10000000000.0 -0.510137170481 approx_equal ERROR 0 approx_equal ERROR Traceback (most recent call last): File "/Users/luc/Developer/cctbx/cctbx_sources/cctbx/cctbx/ regression/tst_twin_target.py", line 540, in <module> run() File "/Users/luc/Developer/cctbx/cctbx_sources/cctbx/cctbx/ regression/tst_twin_target.py", line 488, in run tst_ls_on_i() File "/Users/luc/Developer/cctbx/cctbx_sources/cctbx/cctbx/ regression/tst_twin_target.py", line 95, in tst_ls_on_i assert approx_equal( (fdif_real-old_derivs[0][N_test])/fdif_real, 0, eps=5e-1) AssertionError

2 1

float error
by Alexander.Schiffer＠sanofi-aventis.com 27 Feb '07

27 Feb '07

Dear all, I am trying to detwin a dataset using mmtbx.xtriage but when I run the program with the following command: mmtbx.xtriage data.mtz n_residues=1341 log=detwin.log hklout=data_detwin.mtz symmetry.action=detwin symmetry.twinning_parameters.twin_law=-h-k,k,-l twinning.twinning_parameters.fraction=0.41 I get this message: Attempting to detwin data ------------------------- Detwinning data with: - twin law: -h-k,k,-l BE WARNED! DETWINNING OF DATA DOES NOT SOLVE YOUR TWINNING PROBLEM! PREFERABLY, REFINEMENT SHOULD BE CARRIED OUT AGAINST ORIGINAL DATA ONLY USING A TWIN SPECIFIC TARGET FUNCTION! Traceback (most recent call last): File "/soft/ccp4/cctbx/cctbx_sources/mmtbx/mmtbx/command_line/xtriage.py", line 5, in <module> xtriage.run(command_name="mmtbx.xtriage", args=sys.argv[1:]) File "/soft/ccp4/cctbx/cctbx_sources/mmtbx/mmtbx/scaling/xtriage.py", line 691, in run out=log) File "/soft/ccp4/cctbx/cctbx_sources/mmtbx/mmtbx/scaling/massage_twin_detwin_data.py", line 330, in __init__ self.final_array = detwinner.detwin_it(alpha=self.params.symmetry.twinning_parameters.fraction) File "/soft/ccp4/cctbx/cctbx_sources/mmtbx/mmtbx/scaling/massage_twin_detwin_data.py", line 197, in detwin_it print >> self.out, "Detwinning the data with fraction %3.2f"%(alpha) TypeError: float argument required Any idea? Best regards, Alexander Build tag: 2007_01_24_2154 Red Hat WS 3 Mit freundlichen Grüßen / Best regards / Cordialement Dr. Alexander Schiffer Sanofi-Aventis Deutschland GmbH S&MA Chemical Sciences Germany Structural Biology Industriepark Hoechst Bldg. G877, Room 029 D-65926 Frankfurt am Main t: +49 69 305 24896 f: +49 69 305 80169 w: www.sanofi-aventis.de ************************************************************************************************************************************************************** Sanofi-Aventis Deutschland GmbH · Sitz der Gesellschaft: Frankfurt am Main · Handelsregister: Frankfurt am Main, Abt. B Nr. 40661 Vorsitzender des Aufsichtsrats: Hanspeter Spek - Geschäftsführer: Dr. Heinz-Werner Meier (Vorsitzender), Dr. Matthias Braun, Hervé Gisserot, Dieter Kohl, Prof. Dr. Dr. Werner Kramer, Dr. Martin Siewert **************************************************************************************************************************************************************

2 1

Re: [cctbxbb] fractional coordinate expansion ... share occupancy
by Ralf W. Grosse-Kunstleve 06 Feb '07

06 Feb '07

Hi Andrew, > My question is what is the best method for generating the correct > distribution of atoms at these points. Naively, one would just > randomly select 30% of the shared points to be aluminums and set the > other 70% as silicon. I doubt that this is ideal. Short answer: there is nothing in the cctbx that would help you directly. I have a past as a zeolite person. My understanding was that the 30%/70% occupancy factors are a way to model spatial averages, as that's all you get from the diffraction experiment. How to handle the overlapping atoms depends highly on the context. E.g. in refinement I'd keep the overlapping atoms but constrain the coordinates to be identical. For graphical presentation I'd probably want to find a nice way to present the mixed occupancy, i.e. I wouldn't want to simply draw two crosses on top of each other. What is your context? Cheers, Ralf ____________________________________________________________________________________ Never Miss an Email Stay connected with Yahoo! Mail on your mobile. Get started! http://mobile.yahoo.com/services?promote=mail

1 0

Re: [cctbxbb] fractional coordinate expansion ... share occupancy
by Ralf W. Grosse-Kunstleve 05 Feb '07

05 Feb '07

Hi David, > I have provided an example compound with scatterer code below as you > will note the output has overlapping atoms and the resulting 58 points > are not enough to cover the chemical formula. What do I need to do to > fix this? Sorry, I don't understand the question. Could you please explain what you expect to see as the output? Ralf ____________________________________________________________________________________ Get your own web address. Have a HUGE year through Yahoo! Small Business. http://smallbusiness.yahoo.com/domains/?p=BESTDEAL

2 1

fractional coordinate expansion ... share occupancy
by David A Carr 02 Feb '07

02 Feb '07

Hello All, Thanks to Ralph's modification in January I have had success with cctbx. Now I have another question. Is there a method in the cctbx for dealing with shared occupancy coordinates? I have a compound where in the reduced cell the x,y,z coordinates for two different atoms are the same and a "sof" fraction is given. What is the best practice for handling such points? I have provided an example compound with scatterer code below as you will note the output has overlapping atoms and the resulting 58 points are not enough to cover the chemical formula. What do I need to do to fix this? Thanks, Andrew **************************************** Dr. D. Andrew Carr George Mason University dcarr2(a)gmu.edu ## ## The following example is ## Na9.52 (Al7.2 Si16.8 O48) (H2 O)5.88 ## DAC 1.12.07 ##--------------------------------------- from cctbx import xray from cctbx import crystal from cctbx.array_family import flex def demo(): crystal_symmetry = crystal.symmetry( unit_cell="13.12,13.57,15.81,90.0,143.0,90.000000", space_group_symbol="C 1 2/m 1") scatterers = flex.xray_scatterer() scatterers.append(xray.scatterer(#element = "Si", label="Si1", site=(0.0042,0.2561,0.0965),occupancy=0.7)) scatterers.append(xray.scatterer(#element = "Al", label="Al1", site=(0.0042,0.2561,0.0965),occupancy=0.3)) scatterers.append(xray.scatterer(#element = "Si", label="Si2", site=(0.3442,0.1125,0.3278),occupancy=0.7)) scatterers.append(xray.scatterer(#element = "Al", label="Al2", site=(0.3442,0.1125,0.3278),occupancy=0.3)) scatterers.append(xray.scatterer(#element = "Si", label="Si3", site=(0.3497,0.1193,0.1355),occupancy=0.7)) scatterers.append(xray.scatterer(#element = "Al", label="Al3", site=(0.3497,0.1193,0.1355),occupancy=0.3)) #--------------------------------------------- # Oxygens #--------------------------------------------- scatterers.append(xray.scatterer(#element = "O", label="O1", site=(0,0.3203,0))) scatterers.append(xray.scatterer(#element = "O", label="O2", site=(0.4292,0.1372,0.2924))) scatterers.append(xray.scatterer(#element = "O", label="O3", site=(-.0283,0.3315,0.1491))) scatterers.append(xray.scatterer(#element = "O", label="O4", site=(0.5,0.1347,0.5))) scatterers.append(xray.scatterer(#element = "O", label="O5", site=(0.1993,0.2072,0.2422))) scatterers.append(xray.scatterer(#element = "O", label="O6", site=(0.308,0,0.0850))) scatterers.append(xray.scatterer(#element = "O", label="O7", site=(-.1454,0.1619,0.0205))) scatterers.append(xray.scatterer(#element = "O", label="O8", site=(0.256,0,0.289))) scatterers.append(xray.scatterer(#element = "Na", label="Na", site=(0.7235,0.2619,0.3904),occupancy=1.19)) scatterers.append(xray.scatterer(#element = "O", label="O9", site=(0.663,0.5645,0.7249),occupancy=0.47)) scatterers.append(xray.scatterer(#element = "O", label="O10", site=(0.304,0,0.686),occupancy=0.53)) icsd_structure = xray.structure( crystal_symmetry=crystal_symmetry, scatterers=scatterers) icsd_structure.show_summary().show_scatterers() print icsd_structure.show_distances(distance_cutoff=2.5) print primitive_structure = icsd_structure.primitive_setting() primitive_structure.show_summary().show_scatterers() print p1_structure = primitive_structure.expand_to_p1() p1_structure.show_summary().show_scatterers() print print "OK" if (__name__ == "__main__"): demo()

1 0

Re: [cctbxbb] set_use_u_iso and set_use_u_aniso
by Ralf W. Grosse-Kunstleve 08 Jan '07

08 Jan '07

> > What are those methods of xray.scatterer for? > > And what are their relation to set_grad_u_iso and set_grad_u_aniso? You can use any combination of u_iso and u_star (anisotropic). We had to make this generalization to accomodate TLS refinement. I.e. there are four combinations, use_u_iso/use_u_aniso false/false, true/false, false/true, true/true. The total Debye-Waller factor is the product dw_iso * dw_aniso. If use_u_iso is false, dw_iso = 1, similarly dw_aniso = 1 if use_u_ansio is false. Unfortunately I didn't find the time yet to eliminate the anisotropic_flag, which is redundant. The main problem is that phaser needs to be changed in many places in order to get rid of the flag. The grad_u_iso and grad_u_ansio control the computation of the gradients. The flags are independent of corresponding use flags. __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

1 0

set_use_u_iso and set_use_u_aniso
by Luc Bourhis 08 Jan '07

08 Jan '07

What are those methods of xray.scatterer for? Luc

1 1

Re: [cctbxbb] Error in fractional coordinates returned by cctbx
by Ralf W. Grosse-Kunstleve 29 Dec '06

29 Dec '06

> We could add .mod_positive() and .mod_short() methods to flex.xray_scatterer, and > then the corresponding two-liners .sites_mod_positive() and .sites_mod_short() to > xray.structure. I'll put it on my to-do list. Done: http://cci.lbl.gov/cctbx_build/all.html Build 2006_12_28_2147 or later includes the new methods. E.g.: p1_structure = structure.expand_to_p1().sites_mod_positive() HTH, Ralf __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

2 1

Re: [cctbxbb] Error in fractional coordinates returned by cctbx
by Ralf W. Grosse-Kunstleve 29 Dec '06

29 Dec '06

> flex.fmod_positive( coordinates, 1 ) > > on the fractional coordinates on your sites. > > Ralf, is there a better way? No, Peter's Python loop over the scatterers is currently the best solution. We could add .mod_positive() and .mod_short() methods to flex.xray_scatterer, and then the corresponding two-liners .sites_mod_positive() and .sites_mod_short() to xray.structure. I'll put it on my to-do list. Cheers, Ralf __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

1 0

Error in fractional coordinates returned by cctbx
by David A Carr 28 Dec '06

28 Dec '06

Hello, I have encountered the following inconsistency when working with cctbx. The compound I am working with is as follows. #Silicon Dioxide Quartz low. #Si O2 #ICSD Coll Number 62410 label="Si", site=(0.4643,0.0,0.6667) label="O", site=(0.411,0.2773,0.7783) unit_cell="4.8430, 4.8430, 5.3480, 90.0, 90.0, 120.0", space_group_symbol="P 32 2 1" The error is can be seen in the below. In the second and third Si lines the z value is not a fraction of the unit. Similarly two of the oxygens are placed outside the unit cell. Why does cctbx place these outside of the unit cell? Number of scatterers: 9 At special positions: 0 Unit cell: (4.843, 4.843, 5.348, 90, 90, 120) Space group: P 1 (No. 1) Label, Scattering, Multiplicity, Coordinates, Occupancy, Uiso Si1 Si 1 ( 0.4643 0.0000 0.6667) 1.00 0.0000 Si1 Si 1 ( 0.0000 0.4643 1.3333) 1.00 0.0000 Si1 Si 1 (-0.4643 -0.4643 1.0000) 1.00 0.0000 O1 O 1 ( 0.4110 0.2773 0.7783) 1.00 0.0000 O1 O 1 (-0.2773 0.1337 1.4450) 1.00 0.0000 O1 O 1 (-0.1337 -0.4110 1.1116) 1.00 0.0000 O1 O 1 ( 0.1337 -0.2773 -0.4450) 1.00 0.0000 O1 O 1 (-0.4110 -0.1337 -0.1116) 1.00 0.0000 O1 O 1 ( 0.2773 0.4110 -0.7783) 1.00 0.0000 Below shows the expected fractional coordinates as expected. Based on the space group P 32 2 1. 0.4643 0.0000 0.6667 0.0000 0.4643 0.3333 -0.4643 -0.4643 0.0000 0.4110 0.2773 0.7783 -0.2773 0.1337 0.4450 -0.1337 -0.4110 0.1116 0.1337 -0.2773 -0.4450 -0.4110 -0.1337 -0.1116 0.2773 0.4110 -0.7783 Any insight into this would be of help. I am hoping to use cctbx as a part of our current software development project. Thanks, D.Andrew Carr dcarr2(a)gmu.edu # Below is the code that generates the error # posted above. #--------------------------------------------------- from cctbx import xray from cctbx import crystal from cctbx.array_family import flex def demo(): crystal_symmetry = crystal.symmetry( unit_cell="4.8430, 4.8430, 5.3480, 90.0, 90.0, 120.0", space_group_symbol="P 32 2 1") scatterers = flex.xray_scatterer() scatterers.append(xray.scatterer( label="Si", site=(0.4643,0.0,0.6667))) scatterers.append(xray.scatterer( label="O", site=(0.411,0.2773,0.7783))) icsd_structure = xray.structure( crystal_symmetry=crystal_symmetry, scatterers=scatterers) print "ICSD Structure" icsd_structure.show_summary().show_scatterers() print icsd_structure.show_distances(distance_cutoff=2.5) print "Primitive" primitive_structure = icsd_structure.primitive_setting() primitive_structure.show_summary().show_scatterers() print p1_structure = primitive_structure.expand_to_p1() p1_structure.show_summary().show_scatterers() print print "Expanded to P1" icsd_structure.expand_to_p1().show_summary().show_scatterers() print "OK" if (__name__ == "__main__"): demo()

3 8