Dear Ed,
It turns out that the format of the structure file I sent to the PDB guys is
not
right. I found out this when running phenix.model_vs_data. After I resent the
structrue file, sfcheck gives something OK, which is,
High_Res Low_Res Compl Num_Ref R_obs R_work R_free Corr(Fo-Fc)
Reported PHENIX 1.695 36.657 95.51 46167 0.1488 0.1474 0.1742
N/A
SFCHECK without TLS 1.69 39.42 95.9 45158 0.1604 0.162 0.183
0.9546
REFMAC without TLS 1.694 39.426 95.428 46172 0.148 0.1464 0.1736
0.964
PHENIX 1.69 36.66 96.0 46167 N/A 0.1475 0.1741
N/A
Thank you all for your suggestions,
Best wishes,
Vinson
________________________________
发件人： Ed Pozharski <epozh...@umaryland.edu>
收件人： CCP4BB@JISCMAIL.AC.UK
发送日期： 2010/10/23 (周六) 12:07:48 上午
主 题： Re: [ccp4bb] R-free flag problem
I must note that it is also interesting to know why the sfcheck produced
the R-factor that is twice as high. Of course, the R/Rfree are not
supposed to match precisely (as I recall, sfcheck uses Babinet bulk
solvent correction which is presumably inferior to the mask-based
approach), but 28.6 versus 14.7? Something may be seriously wrong.
On Fri, 2010-10-22 at 09:37 +0200, Tim Gruene wrote:
> Dear Vinson,
>
> I would still be interested about what was the cause of the missing R-free
>flags
> and how you resolved it (I hope I did not miss an email).
>
> Kind regards, Tim
>
> On Fri, Oct 22, 2010 at 02:33:03PM +0800, Vinson LIANG wrote:
> >
> >
> > Dear all,
> >
> >
> >
> >
> > Thank you very much for your suggestions.
> >
> > I try Pavel's easy method. Hopefully, the result from phenix.model_vs_data
>would be enough for PDB.
> >
> >
> > Best wishes,
> >
> > Vinson
> >
> >
> >
> >
> >
> >
> >
> >
> > 发件人： Pavel Afonine <pafon...@lbl.gov>
> > 收件人： PHENIX user mailing list <pheni...@phenix-online.org>
> > 抄 送： Vinson LIANG <lwg_conrad_1...@yahoo.com.cn>
> > 发送日期： 2010/10/22 (周五) 12:44:58 上午
> > 主 题： Re: [phenixbb] R-free problem with phenix.refine
> >
> > Hi,
> >
> >
> >
> >
> > Recently I deposited a structure to PDB. However, I was told
> > >> The R value of SF check are different from your reported value.
> > >>
> > >> R work in SF check = 0.2859
> > >> your reported R work = 0.1474
> > >> Also there are no R free flags in your sending files, but you reported R
> > >> free value in your all deposition.
> > well, SF check is not the best tool to check things - it's not just my only
>vision, see lots of conversations on this topic on bb.
> >
> > Just run
> >
> > phenix.model_vs_data model.pdb data.hkl
> >
> > and if output statistics matches the one you deposited to PDB, then you are
>100% good. If it doesn't match then there is something to worry about.
> >
> >
> >
> > So, I rerun my deposited file with phenix.refine, and it gave a
>R-work=0.1475, R-free=0.1742.
>
> >
> > Sounds good.
> >
> >
> >
> > However, when I tried to run the same file with refmac5, error came with
>"Free R label has not been set". I guess phenix.refine gives Rree-R label that
>other program could not recognize.
>
> >
> > No surprises. Jumping between programs is tricky. You have to account for a
>lot of things to match things that different programs expect differently.
> >
> >
> >
> > Could anyone please help me to solve this problem?
> > Step-by-step:
> >
> > 1) Run
> > phenix.model_vs_data model.pdb data.hkl
> >
> > 2) if output statistics matches the one you deposited to PDB, then you are
>100% good. Come back to PDB and say this.
>
> >
> > 3) If it doesn't match then there is something that you need to fix. If you
>can't fix or have no idea what's wrong, you can send me the data and model and
>I
>will have a look.
> >
> > Pavel.
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
--
Edwin Pozharski, PhD, Assistant Professor
University of Maryland, Baltimore
----------------------------------------------
When the Way is forgotten duty and justice appear;
Then knowledge and wisdom are born along with hypocrisy.
When harmonious relationships dissolve then respect and devotion arise;
When a nation falls to chaos then loyalty and patriotism are born.
------------------------------ / Lao Tse /
