[phenixbb] Few questions

Pavel Afonine pafonine at lbl.gov
Thu Mar 15 07:57:14 PDT 2007

Hi John,

> Is it possible to set the maximum B-factor limit  during adp 
> refinement ( like  in CNS... 5-100)

In phenix.refine you can reset B-factors (selected or all) before 
1) you can ransomize them;
2) set to a value;
3) set to Wilson B.

To do so, look modify_start_model scope:

refinement.modify_start_model {
  adp {
    randomize = None
    set_b_iso_to = None
    set_biso_to_wilson_b = None
    convert_to_isotropic = False
    shift_b_iso_by = None
  selection = None
  write_modified = None


phenix.refine data.mtz model.pdb modify_start_model.adp.set_b_iso_to=20 
modify_start_model.selection="chain A and resid 10-20"

This will set B=10 for the residues from 10 to 20 in chain A.

Make sure to refine your model after such manipulation. Also, large B 
may indicate either lack of density or potential big movements so you 
may want to use TLS model.

> Regarding rmsd question.. after model building and refinement ...i 
> took the model and ran only xtal_pdbsubmission  in cns. And that gave 
> the difference rmsds (phenix 0.007   & cns 0.0218) thats why i got 
> confused.

If you prepare this in CNS, you can manually replace the values. Also, 
make sure to put PHENIX (phenix.refine) as a program used.

Let me know if you have any questions!


More information about the phenixbb mailing list