[phenixbb] R, Rfree, and wxu_scale

Jianghai Zhu jzhu at cbr.med.harvard.edu
Sat Aug 4 19:43:51 PDT 2007


Yes, yes.  But before one admits the model with the higher Rfree is  
the more accurate model, one must make sure the refinement protocol  
is correct.

Jianghai

+++++++++++++++++++++++++++++++
Jianghai Zhu, Ph.D
CBR Institute for Biomedical Research
Department of Pathology
Harvard Medical School
200 Longwood Ave., Boston, MA 02115
Ph: 617-278-3211
Fx: 618-278-3232
+++++++++++++++++++++++++++++++



On Aug 4, 2007, at 10:06 PM, Mayer, Mark (NIH/NICHD) [E] wrote:

> Forgive me for saying this but the object of refinement is to build  
> the most accurate model, not to obtain the lowest Rvalue,
> which will vary with weights applied to diffraction data versus  
> constraints. There have been multiple posts about this on
> crystallographic BBs in the last few months ...
>
>
> -----Original Message-----
> From: Jianghai Zhu [mailto:jzhu at cbr.med.harvard.edu]
> Sent: Sat 8/4/2007 9:19 PM
> To: PHENIX user mailing list
> Subject: [phenixbb] R, Rfree, and wxu_scale
>
> Hi,
>
> Just finished a round of refinement.  Rfree went up about 1% and R
> went down a bit.  I am wondering what happened here.  Also Rfree
> always went down after TLS refinement but always went up a lot more
> after the individual ADP refinement.  Is there anything I can do
> here?  Should I use weaker B restraints?  I tried to assign a
> wxu_scale value.  But phenix.refine automatically adjust it to 1.81
> every time.
>
> ---------------------------------------------------------------------- 
> --
> R-factors, x-ray target values and norm of gradient of x-ray target
> stage     r-work r-free  xray_target_w  xray_target_t
>     0    :  0.3982 0.5701   6.156859e+00   6.238509e+00
>     1_bss:  0.2008 0.2494   5.949117e+00   6.088855e+00
>     1_xyz:  0.2030 0.2497   5.953829e+00   6.091696e+00
>     1_adp:  0.1998 0.2559   5.952833e+00   6.099340e+00
>     2_bss:  0.1996 0.2557   5.952517e+00   6.099196e+00
>     2_xyz:  0.1984 0.2576   5.953246e+00   6.104085e+00
>     2_adp:  0.1974 0.2578   5.950434e+00   6.102336e+00
>     3_bss:  0.1972 0.2574   5.950333e+00   6.102222e+00
>     3_xyz:  0.1997 0.2576   5.956524e+00   6.103647e+00
>     3_adp:  0.1987 0.2590   5.955554e+00   6.104434e+00
>     3_bss:  0.1986 0.2588   5.955445e+00   6.104227e+00
> ---------------------------------------------------------------------- 
> --
> Weights for target T = Exray * wxc * wxc_scale + Echem * wc and
> angles between gradient vectors, eg. (d_Exray/d_sites, d_Echem/ 
> d_sites)
> stage              wxc          wxu  wxc_sc  wxu_sc /_gxc,gc /_gxu,gu
>     0    :    5.5611e-01   7.2249e-01   0.050   1.810   96.334   
> 101.736
>     1_bss:    5.5611e-01   7.2249e-01   0.050   1.810   96.334   
> 101.736
>     1_xyz:    4.5001e-01   6.1023e-01   0.050   1.810   96.591   
> 102.016
>     1_adp:    4.5001e-01   6.1023e-01   0.050   1.810   96.591   
> 102.016
>     2_bss:    4.5001e-01   6.1023e-01   0.050   1.810   96.591   
> 102.016
>     2_xyz:    5.1326e-01   6.5356e-01   0.050   1.810  112.802   
> 141.393
>     2_adp:    5.1326e-01   6.5356e-01   0.050   1.810  112.802   
> 141.393
>     3_bss:    5.1326e-01   6.5356e-01   0.050   1.810  112.802   
> 141.393
>     3_xyz:    3.3568e-01   7.0695e-01   0.050   1.810  111.975   
> 135.760
>     3_adp:    3.3568e-01   7.0695e-01   0.050   1.810  111.975   
> 135.760
>     3_bss:    3.3568e-01   7.0695e-01   0.050   1.810  111.975   
> 135.760
> ---------------------------------------------------------------------- 
> --
>
> Jianghai
>
> +++++++++++++++++++++++++++++++
> Jianghai Zhu, Ph.D
> CBR Institute for Biomedical Research
> Department of Pathology
> Harvard Medical School
> 200 Longwood Ave., Boston, MA 02115
> Ph: 617-278-3211
> Fx: 618-278-3232
> +++++++++++++++++++++++++++++++
>
>
>
>
> <winmail.dat>
> _______________________________________________
> phenixbb mailing list
> phenixbb at phenix-online.org
> http://www.phenix-online.org/mailman/listinfo/phenixbb

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://phenix-online.org/pipermail/phenixbb/attachments/20070804/be5b1693/attachment-0003.htm>


More information about the phenixbb mailing list