Thank you Tim and Pavel. Sorry, just to follow up: For the example I gave R /Rfree difference seems high. Can this still be down to statistical fluctuations? I am just trying to get a feel when we can simply put it down to statistical fluctuations and when its a concern. On Mon, Jan 26, 2015 at 4:26 PM, Tim Gruene wrote:

Dear George,

you may also want to take into account that the variation of Rfree is estimated as Rfree^2/k, where k is the number of reflections in the shell. If you only have e.g. 100 reflections there, this might simply a statistical fluctuation, as far as I understand.

Cheers, Tim

On 01/26/2015 10:05 PM, George Devaniranjan wrote:

...

Hi,

I calculated/validated several PDB structures in Phenix.Validation (GUI)

The R/R-free (overall) is

0.1271/0.1348

Further down I looked at the R/R-free for the highest resolution shell.

Its 0.2538/0.2446

I am curious why the R-free for the *highest resolution shell* seems to be LOWER than R (work).

Am I reading the information wrong?

PDB: 3QL9

Thank you, George

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

-- Dr Tim Gruene Institut fuer anorganische Chemie Tammannstr. 4 D-37077 Goettingen

GPG Key ID = A46BEE1A

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

Dear Pavel, Once the Ramachandran outliers (and the rotamer outliers) was corrected by coot, another round of Phenix refine usually cause the rising of the Rwork and Rfree. Is any way to avoid this rising? After this Rwork and Rfree rising, can I continue the refinement with the Rwork and Rfree rising PDB and MTZ files to try to lower the Rwork and Rfree? Smith On Tuesday, January 27, 2015 3:28 PM, Pavel Afonine wrote: George, what exactly you are trying to achieve by making sense of Rfree and its relation to Rwork in resolution shells? Just curious.. Pavel On 1/26/15 1:50 PM, George Devaniranjan wrote: Thank you Tim and Pavel. Sorry, just to follow up: For the example I gave R /Rfree difference seems high. Can this still be down to statistical fluctuations? I am just trying to get a feel when we can simply put it down to statistical fluctuations and when its a concern. On Mon, Jan 26, 2015 at 4:26 PM, Tim Gruene wrote: Dear George, you may also want to take into account that the variation of Rfree is estimated as Rfree^2/k, where k is the number of reflections in the shell. If you only have e.g. 100 reflections there, this might simply a statistical fluctuation, as far as I understand. Cheers, Tim On 01/26/2015 10:05 PM, George Devaniranjan wrote:

Hi,

I calculated/validated several PDB structures in Phenix.Validation (GUI)

The R/R-free (overall) is

0.1271/0.1348

Further down I looked at the R/R-free for the highest resolution shell.

Its 0.2538/0.2446

I am curious why the R-free for the *highest resolution shell* seems to be LOWER than R (work).

Am I reading the information wrong?

PDB: 3QL9

Thank you, George

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

-- Dr Tim Gruene Institut fuer anorganische Chemie Tammannstr. 4 D-37077 Goettingen GPG Key ID = A46BEE1A _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb