[phenixbb] does phenix.refine ignore F-obs, SIGF-obs?

Pavel Afonine PAfonine at lbl.gov
Thu Mar 11 13:34:12 PST 2010


Hi Scott,

phenix.refine removes only zero or strictly negative values. ONLY. NOthing else!  It will not remove a reflection like Fobs=0.0001, regardless sigma.

 REMARK   3   MIN(FOBS/SIGMA_FOBS)              : 0.03

means that this is what in your data file, but it does NOT mean it removed any positive Fobs.

Pavel.

PS> I'm not in a convenient place to look into it carefully, but if you send me the data I will have a look in a few days when I come back to Berkeley.



----- Original Message -----
From: Scott Classen <SClassen at lbl.gov>
Date: Thursday, March 11, 2010 4:27 pm
Subject: Re: [phenixbb] does phenix.refine ignore F-obs, SIGF-obs?
To: PHENIX user mailing list <phenixbb at phenix-online.org>

> 
> Here are some excerpts from the phenix.refine log file:
> 
> Resolution range: 44.4024 1.8604
> Completeness in resolution range: 0.995991
> Completeness with d_max=infinity: 0.995862
> 
> Number of F-obs in resolution range:                   92411
> Number of F-obs <= 0:                                  3322
> Refinement resolution range: d_max =  44.4024
>                             d_min =   1.8604
> 
> 
> Fobs statistics after all cutoffs applied:
> 
> Miller array info: None
> Observation type: xray.amplitude
> Type of data: double, size=89089
> Type of sigmas: double, size=89089
> Number of Miller indices: 89089
> Anomalous flag: False
> Systematic absences: 0
> Centric reflections: 11672
> Resolution range: 44.4024 1.8604
> Completeness in resolution range: 0.960187
> Completeness with d_max=infinity: 0.960063
> 
> 
> 
> This implies to me that phenix.refine is throwing out my very weak 
> reflections. It appears that completeness goes from 99.6% to 96.0%
> 
> Then further down in the log file I see:
> 
> ============================== Outliers rejection 
> =============================
> Distribution of F-obs values before outliers rejection:
>     0.126 - 28.714     : 65519
>    28.714 - 57.302     : 13850
>    57.302 - 85.890     : 5625
>    85.890 - 114.478    : 2515
>   114.478 - 143.066    : 991
>   143.066 - 171.654    : 393
>   171.654 - 200.242    : 128
>   200.242 - 228.830    : 49
>   228.830 - 257.418    : 12
>   257.418 - 286.006    : 7
> 
> basic_wilson_outliers    = 0
> extreme_wilson_outliers  = 0
> beamstop_shadow_outliers = 0
> model_based_outliers     = 1
> total                    = 1
> 
> Distribution of F-obs values after outliers rejection:
>     0.126 - 28.714     : 65519
>    28.714 - 57.302     : 13850
>    57.302 - 85.890     : 5624
>    85.890 - 114.478    : 2515
>   114.478 - 143.066    : 991
>   143.066 - 171.654    : 393
>   171.654 - 200.242    : 128
>   200.242 - 228.830    : 49
>   228.830 - 257.418    : 12
>   257.418 - 286.006    : 7
> 
> 
> which makes it look like nothing is being rejected?
> 
> So what is it?
> 
> in the header of the resulting pdb file:
> 
> REMARK   3  DATA USED IN REFINEMENT.
> REMARK   3   RESOLUTION RANGE HIGH (ANGSTROMS) : 1.860   
> REMARK   3   RESOLUTION RANGE LOW  (ANGSTROMS) : 44.402  
> REMARK   3   MIN(FOBS/SIGMA_FOBS)              : 0.03  
> REMARK   3   COMPLETENESS FOR RANGE        (%) : 96.02 
> REMARK   3   NUMBER OF REFLECTIONS             : 89088     
> 
> 
> which implies that indeed some reflections are being rejected.
> 
> My hunch is that if F-obs = SIGF-obs then the reflection is rejected.
> 
> Thanks,
> Scott
> _______________________________________________
> phenixbb mailing list
> phenixbb at phenix-online.org
> http://phenix-online.org/mailman/listinfo/phenixbb
> 



More information about the phenixbb mailing list