Re: [ccp4bb]: B-factor correction at low resolutin in cns

[<tsr@duke.edu>]

***  For details on how to be removed from this list visit the  ***
***          CCP4 home page http://www.ccp4.ac.uk         ***

On Sat, 28 Sep 2002, leesea wrote:

> ***  For details on how to be removed from this list visit the  ***
> ***          CCP4 home page http://www.ccp4.ac.uk         ***
>
> Dear all, When refining a structure at 3.5 angstrom with CNS, I get
> zero/negative B values after minimizing or annealing. I found it's the
> initial B-factor correction that lead to such change. I have to choose
> 'no correction'and give a overall B value of 20 to keep things normal,
> but I am not sure whether it is proper. However, the present model fits
> the electron density well (R/Rfree = 0.29/0.33).So can anyone tell me
> what's the main factor that causes the failure of B-factor correction?
> If it's a result of low resolution, what's the best overall B value
> should we choose? Thanks for your help.
>
> 　　　　　　　　　　　　　　leesea@moon.ibp.ac.cn
> 　　　　　　　　　　　　　　　　　2002-09-28
>
>
>
>
Leesea,

I've had the same problem with datasets at ~3.5A resolution.  In CNS, try
switching the B-factor correction from 'anisotropic' to
'anisotropic-fixed-isotropic' so that the B-factor correction is applied
only to the data.  This should help to keep the B-factors in the model
realistic.  In terms of what overall B factor to choose, the low(er)
diffraction limit may be indicative of more disorder in the crystal.  An
overall B-factor of 40-50 may be more appropriate in this case.  Best of
luck.

T. Scott Reid
tsr@duke.edu