Are you sure that the mtz file you are refining against has spacegroup
C2221 in the header? Phaser can test C222 and C2221 but when you start
refinement you need to be sure the data header is correct.
Another possible bug - is the model elongated in any way - they
sometimes need surface loops clipped out. I would always recommend
running chainsaw before you start the search, then looking at the model
to see if there are extruberances.
eleanor
On 11/29/2010 01:16 PM, xiuwen zhang wrote:
> Dear colleages,
>
> Currently I met problem in a MR case and hope somebody could advise me.
>
? > The crystals belong to sg C2221 with the cell parameters a=59A,
b=302A,
> c=97A, beta=92.3. The searching model shares 43% identities (300aa) with
> target protein. When searching in PHASER program, only one clear solution
> could be found with high Z-score: RZF=9.8 and TZF=24.2. However, though the
> density is interpretative, the solution coordinate could not fit in the
> density (Rfactor high to 54%). Moreover, as suggested by matthews_coff,
> there should be at least 2 molecules in one ASU. But I could not find the
> other molecules in the ASU in this case (though the other solution are
> defined with high Z-score RZF=7.8 and TZF=59, this solution has 115 clashes
> with the first one). I also find that this solution coordinate occupied some
> "layers" and most of the unit cell are lack of density. As
> Phenix.xtrage suggested the data is not twinned, I was totally confused by
> this situation.
>
> I will appreciate if you could share your opinions and experience on
> similar case with me.
>
> Best Regards,
> Xiuwen
>
|