Re: [ccp4bb]: about REFMAC5

[Garib Murshudov <garib@ysbl.york.ac.uk>]

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

GUAN wrote:

> ***  For details on how to be removed from this list visit the  ***
> ***    CCP4 home page http://www.dl.ac.uk/CCP/CCP4/main.html    ***
>
> I use CNS to refine a structure to R=16.99% and Rf=18.92% with
> a 1.4 A data set, and use REFMAC5 to refine anisotropic B factors,
> as suggested by many experts on CCP4 BB. The result showed a R
> of 13.11% and Rf of 16.50%.

Usually I do isotropic refinement and anisotrpic refinements if free R
for aniso is
less than 1% lower than that for iso it is o.k. (you should use same
program for all options
and compare results. Statistics in different programs my be calculated
by different means hence
difficult to compare)

Now you can try 3 options:

individual isotropic refinement
TLS + isotropic refinement (each subunit or domain as one rigid group)
individual anisotropic refinement
and compare resuts

>
> But an expert here said that the number of my reflections is not
> enough for this refinement, so the result was some meaningless.
> I have 19255 reflections, 132 residues and 146 waters in the A.U.,
> (total 1154 atoms). Each atom has 9 parameters to be refined
> (3 coordinate parameters and 6 anisotropic B factor parameters),
> so the ratio of reflection over parameters is about 2, not very
> high.

In this case it looks o.k. to me observ/parameters is around 2. You
should remember that
restraints are also used.

Garib

--
------------------------------------------------------------------
Garib N. Murshudov, Chemistry Department, University of York, U.K.
Tel: Home +44 (1904) 42 62 87, work:  +44 (1904) 43 25 65
------------------------------------------------------------------