Hi, Pavel.
Thanks for the quick reply. I downloaded the files, and I got the result you
predicted - all the occupancies looked neat and tidy as expected. Since you
used those nice occupancies to generate perfect data, which was then refined
against, that's what one would expect, right? So I am still left with the
issue of the occupancies going nuts in my real refinement. Or am I missing
something?
Regards,
Anna
On Tue, Feb 10, 2009 at 4:43 PM, Pavel Afonine
Hi Anna,
to find out what's going on I just did the following:
1) I copied the lines below into "answer.pdb" file and I re-set the occupancies so they all make sense (add up to 1);
2) Using "answer.pdb" file I simulated two datasets: X-ray and neutron: "data.mtz". What I did is just I computed X-ray and neutron Fcalc, and called their absolute values as F-obs-xray and F-obs-neutron, and stored them in "data.mtz".
3) Then I scrambled the model in "answer.pdb" by randomizing the occupancies of all atoms. This file containing the scrambled models is called "start.pdb".
4) Finally, I took "data.mtz" and "start.pdb" files and ran joint XN refinement in phenix.refine. I did only occupancy refinement. As the refinement result I was expecting to get the PDB file with refined structure where all the occupancies are exactly the same as in "answer.pdb". The command I used is:
phenix.refine data.mtz start.pdb params --overwrite
5) Now, open two files "answer.pdb" and "refinement_result_001.pdb", and voila! All the occupancies are as expected.
All the input and output files discussed above are located here: http://cci.lbl.gov/~afonine/for_AnneG/http://cci.lbl.gov/%7Eafonine/for_AnneG/
All the above data and files manipulations were done using "phenix.pdbtools" and "phenix.refine".
Could you reproduce the refinement I just did ("step 4") using your version of PHENIX ? If not, then please update to the latest PHENIX (http://www.phenix-online.org/download/) and try again. If not successful even then, I will send you the latest development version.
Please let me know if you have any questions or problems!
Pavel.
On 2/10/2009 11:04 AM, Anna Gardberg wrote:
I am doing joint refinement of a perdeuterated protein, and I am having a problem with alternate conformations - the deuterium occupancies don't match the overall occupancies for their conformation:
ATOM 335 N APRO A 19 13.154 0.897 15.243 0.38 8.64 N ATOM 336 CA APRO A 19 12.556 0.209 16.391 0.38 7.61 C ATOM 337 CB APRO A 19 13.743 0.043 17.350 0.38 12.47 C ATOM 338 CG APRO A 19 14.713 1.102 16.926 0.38 9.70 C ATOM 339 CD APRO A 19 14.594 1.126 15.443 0.38 13.59 C ATOM 340 C APRO A 19 11.414 0.974 17.057 0.38 13.42 C ATOM 341 O APRO A 19 10.559 0.334 17.671 0.38 15.65 O ATOM 342 DA APRO A 19 12.236 -0.672 16.112 0.48 5.78 D ATOM 343 DB2APRO A 19 13.454 0.173 18.268 0.43 15.49 D ATOM 344 DB3APRO A 19 14.128 -0.841 17.229 0.52 10.36 D ATOM 345 DG2APRO A 19 14.460 1.959 17.300 0.60 13.82 D ATOM 346 DG3APRO A 19 15.610 0.852 17.197 0.30 14.04 D ATOM 347 DD2APRO A 19 14.850 1.996 15.097 0.48 12.84 D ATOM 348 DD3APRO A 19 15.119 0.413 15.046 0.40 8.63 D ATOM 349 N BPRO A 19 13.163 0.886 15.240 0.62 8.58 N ATOM 350 CA BPRO A 19 12.606 0.252 16.438 0.62 7.70 C ATOM 351 CB BPRO A 19 13.777 0.307 17.423 0.62 11.89 C ATOM 352 CG BPRO A 19 14.973 0.224 16.548 0.62 11.13 C ATOM 353 CD BPRO A 19 14.616 1.094 15.375 0.62 13.76 C ATOM 354 C BPRO A 19 11.399 0.972 17.012 0.62 13.34 C ATOM 355 O BPRO A 19 10.491 0.304 17.514 0.62 15.31 O ATOM 356 DA BPRO A 19 12.371 -0.680 16.253 0.23 7.26 D ATOM 357 DB2BPRO A 19 13.760 1.148 17.905 0.72 12.42 D ATOM 358 DB3BPRO A 19 13.734 -0.446 18.034 0.19 14.30 D ATOM 359 DG2BPRO A 19 15.759 0.552 17.012 0.87 14.63 D ATOM 360 DG3BPRO A 19 15.094 -0.695 16.264 0.40 9.68 D ATOM 361 DD2BPRO A 19 14.807 2.022 15.580 0.43 11.93 D ATOM 362 DD3BPRO A 19 15.083 0.798 14.578 0.31 8.63 D
I am using these refinement options: refine { strategy = *individual_sites *rigid_body *individual_adp group_adp tls \ *occupancies group_anomalous sites { individual = None rigid_body = None } occupancies { individual = None remove_selection = None constrained_group { selection = None } } }
and hydrogens { refine = *individual riding refine_adp = one_b_per_residue one_b_per_molecule *individual refine_occupancies = one_q_per_residue *one_q_per_molecule individual contribute_to_f_calc = True } What refinement options should I be using to keep the D occupancies the same as the rest of their conformation? Hope you can help!
Regards, Anna
phenixbb mailing list [email protected] http://www.phenix-online.org/mailman/listinfo/phenixbb