Re: [ccp4bb] Amazing B-factor

- Protein crystallography

Main steps:

- Protein purification
- Crystallisation

Special:

- Programs for crystallography
- X-ray detectors

Basic tutorials:

   - Chemistry
   - Protein
   - Peptide
   - Amino Acids

Xtal community:

- CCP4BB

CCP4bb navigation

CCP4bb <-- 1999 <-- November 1999 <-- 30 November 1999

Previous message:
Subject: Re: coupling between occupancy and b-values in refinement
From: Dale Tronrud det102 {- at -} UOXRAY {- dot -} UOREGON {- dot -} EDU
Date: 2009-06-01

Next message:
Subject: Re: BioRad CFX96 for thermofluor/DSF?
From: "Asmussen, Gary" Gary {- dot -} Asmussen {- at -} GENZYME {- dot -} COM
Date: 2009-06-01

Subject: Re: Amazing B-factor
From: Dale Tronrud det102 {- at -} UOXRAY {- dot -} UOREGON {- dot -} EDU
Date: 2009-06-01

Dear Jain,

The Wilson plot is based on the assumption that you can model your
structure with a collection of atoms randomly distributed in the unit
cell. This model is very poor at low resolution, say, 4.5A and lower.
Since the Wilson B depends on the change in scattering as a function of
resolution, a Wilson B calculated for 3A data will be very unreliable,
being based only on the data from 4.5A to 3A.

Once you begin refinement you have a proper model of your crystal
and all the low resolution data can be used to estimate the average
B. At this point the Wilson B should be ignored, it was only a stepping
stone to get you here.

A Wilson B of 50 to 60A^2 for a 3A data set collected on frozen crystals
at a synchrotron (I'm making an assumption here.) is quite surprising.
137A^2 seems much more likely to me. I'd trust CNS's results.

The overall anisotropic correction should be fine at this resolution
too.

Dale Tronrud

Jian Wu wrote:
> Dear all,
>
> Recently we have collected one set of data which is processed to 2.9A
> and 3.0A and the Wilson-B values are 50.1 and 59.1, respectively. As for
> the completeness of the highest shell is only 66% in 2.9A (80% in 3.0A),
> we use the dataset with 3.0A for phasing and refinement.
>
> Everything is OK during phasing. Routinely, we use CNS (rigid, minimize
> and bgroup) for primary refinement and the R and freeR values go to 31.1
> and 34.1. Surprisingly, the B-factor is 137.2. I have modified the
> B-factor to a fixed value (59.1 A^2) in the searching model coordinate
> file, however it changed to 127.6 after 'rigid' and subsequently to
> 131.9 after 'minimize'. In all the scripts above I used the default
> option "anisotrpic" at "overall B-factor correction". I don't know
> whether it is right to use this option at low resolution, so I have
> tried the other two options "no" and "isotropic". In the 1st choice, the
> B-factor is seemingly reasonable (55~59) but the R and freeR are very
> high (45.9 and 51.9). In the 2nd choice, the B-factor, R and freeR
> values are between those of the other two options (111.3, 40.6, and
> 45.3). I want to know what cause these surprising changes especially for
> B-factor value.
>
> Any suggestion is appreciated.
>
> Jian Wu
>
> --
> Jian Wu
>
> Ph.D. Student
> Institute of Biochemistry and Cell Biology
> Shanghai Institutes for Biological Sciences
> Chinese Academy of Sciences (CAS)
> Tel: 0086-21-54921117
> Email: protease@gmail.com
>

CCP4bb navigation

CCP4bb <-- 1999 <-- November 1999 <-- 30 November 1999

Previous message:
Subject: Re: coupling between occupancy and b-values in refinement
From: Dale Tronrud det102 {- at -} UOXRAY {- dot -} UOREGON {- dot -} EDU
Date: 2009-06-01

Next message:
Subject: Re: BioRad CFX96 for thermofluor/DSF?
From: "Asmussen, Gary" Gary {- dot -} Asmussen {- at -} GENZYME {- dot -} COM
Date: 2009-06-01