Dear Pavel

I played with wxu_scale and wu several times. When wu = 0, this unrestrained refinement results in totally unrealistic ADP's. However, if I use wu = 1, wu = 5 or wu = 20, I always get the same results. Is this normal? On the bottom of this e-mail, you can find the target weight settings (for wu = 20).

Kind regards
Elke

target_weights {
    optimize_xyz_weight = False
    optimize_adp_weight = False
    wxc_scale = 0.5
    wxu_scale = 1
    wc = 1
    wu = 20
    fix_wxc = None
    fix_wxu = None
    shake_sites = True
    shake_adp = 10
    regularize_ncycles = 50
    verbose = 1
    wnc_scale = 0.5
    wnu_scale = 1
    rmsd_cutoff_for_gradient_filtering = 3
    force_optimize_weights = False
    weight_selection_criteria {
      bonds_rmsd = None
      angles_rmsd = None
      r_free_minus_r_work = None
      r_free_range_width = None
      mean_diff_b_iso_bonded_fraction = None
      min_diff_b_iso_bonded = None

Elke De Zitter

PhD student, KU Leuven

Biochemistry, Molecular and Structural Biology

Celestijnenlaan 200f - bus 2404, 3001 Leuven

Room 00.84, +32 16 3 26530


Van: Pavel Afonine [[email protected]]
Verzonden: woensdag 25 juni 2014 18:22
Aan: Elke De Zitter; [email protected]
Onderwerp: Re: [phenixbb] Anisotropic B-factor restraints

Hi Elke,

you can make ADP restraints stronger by making X-ray target weight smaller, for example, set wxu_scale weight to something smaller than the default (1.0). You may need to try a few values to see which one works best.

Alternatively, you can make ADP restraints weight, wu, larger than the default (1.0), which is equivalent.

Let me know if you have any questions or need help with this!

Pavel

On 6/25/14, 7:18 AM, Elke De Zitter wrote:
Dear Pavel

Thank you for the quick response. After reading this paper and some of its references, I suppose that Phenix.refine automatically restraint the anisotropic B-factor calculation. However, I want to restraint it more because certain atoms have very long small ellipsoids and atoms in some tyrosine rings are moving in different directions. Is there a way to play with the restraints manually?

Kind regards
Elke

Elke De Zitter

PhD student, KU Leuven

Biochemistry, Molecular and Structural Biology

Celestijnenlaan 200f - bus 2404, 3001 Leuven

Room 00.84, +32 16 3 26530


Van: Pavel Afonine [[email protected]]
Verzonden: dinsdag 24 juni 2014 14:31
Aan: Elke De Zitter; [email protected]
Onderwerp: Re: [phenixbb] Anisotropic B-factor restraints

Hi Elke,

phenix.refine uses similarity restraints on anisotropic ADPs as described here:

http://journals.iucr.org/d/issues/2012/04/00/ba5180/ba5180.pdf

Pavel

On 6/24/14, 5:03 AM, Elke De Zitter wrote:
Dear All

I'm refining a protein structure with a resolution of 1.25 A, using Phenix GUI version 1.9. I am refining the B-factors for the protein atoms anisotropic and those for the water molecules isotropic. How can I restrain the anisotropic B-factors so that bonded atoms have similar ellipsoids (like the SIMU and DELU commands in Shelx making the atoms moving similarly in direction and amount)?

Thanks
Elke

Elke De Zitter

PhD student, KU Leuven

Biochemistry, Molecular and Structural Biology

Celestijnenlaan 200f - bus 2404, 3001 Leuven

Room 00.84, +32 16 3 26530



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