Hola,
If you have done everything you can to improve the refinement (like
trying lower symm spacegroups, leaving out the loops in question,
different kinds of NCS restraints, TLS refinement, etc.), and what you
do see is sufficiently interesting, there should be no problem in
publishing a structure with an Rfree of 0.34.
The Rfree value should not be seen too much as an absolute value, I'm
sure there are structures with an Rfree of 0.2 where clear
improvements can still be made (and which may or may not contain
interesting new information), while there are structures with Rfree of
0.4 which have been carefully refined and for which the new
information content warrants a good publication.
Mark
Mark J. van Raaij
Dpto de Bioquímica, Facultad de Farmacia
Universidad de Santiago
15782 Santiago de Compostela
Spain
http://web.usc.es/~vanraaij/
On 2 Jun 2008, at 23:24, Cui Sheng wrote:
> Dear all:
>
> We have an unrefinable protein complex structure with one of the
> component is just a gfp. The space group shows as a P6122, a
> notorious space group for twining. But I could not detect any sign
> of twinnig using various of softwares. The R_free traps currently at
> 0.34 and no more space to impove. We observed only one of the two
> gfps in ASU is badly fited to the density. In this molecule, there
> are quite some loops in contact to solution looked totally mesy and
> all the parts rest looks just fine. Take those mesy loops in to the
> consideration,account for only 3% residues in ASU we are still
> bitterly puzzled by the fact that our structure is 5% or more away
> from a refined structure.
>
> Native dataset collected at 2.5A
> Molecular replacement gives unique solution using normal gfp as
> searching model.
>
> any good ideas solving our problem?
>
> many thanx in advance
>
> sheng
|