Re: Anisotropic refinement

From: Aaron Oakley (s_aarono@eduserv.its.unimelb.EDU.AU)
Date: Tue Aug 27 1996 - 18:33:52 CDT


On Tue, 27 Aug 1996, Darren Nickerson wrote:

[delitia] 
> I am refining a data set collected on an area detector to 1.3A.  The
> protein contains roughly 400 amino acids, and a cation.  Isotropic
> refinement gives a data to parameter ratio of roughly 5:1 (74161 unique
> reflections).  Data is about 65% complete in highest resolution shell.

Why not collect more complete data?

> I am trying to decide whether anisotropic refinement is justified at this 
> point, . . . R=0.183 Rfree=0.212 (isotropic).  I am refining with shelx, 
> and have observed that refining the entire structure anisotropically 
> results in an approx. 4% decrease in R, but only a 2.25% decrease in 
> Rfree.  This represents over-refinement in my mind, and so I was 

Why is this over-refinement? Your free-R went down!
It is quite common, I think, for the freeR to decrease less
than the Rcryst. No one has perfect data, after all ;-).

Cheers,

Aaron



This archive was generated by hypermail 2b29 : Tue Mar 20 2007 - 03:13:18 CST