Hi Tim, it is a shame that I can't remember that particular paper where I think Randy Read or Axel Brunger have demonstrated that phase error and Rfree are well correlated. So if you get best Rwork and Rfree chances are you also get best map too. All the best, Pavel On 1/12/15 3:54 AM, Tim Gruene wrote:

Dear George,

what are the reasons why are you trying to reduce the R/Rfree instead of producing the best possible map?

Best, Tim

On 01/11/2015 11:24 PM, George D wrote:

...

Hi Phenix users,

I was looking at the phenix-refine manual and it mentions running

optimize_mask=true

optimize_xyz_weight=true

optimize_adp_weight=true

I was wondering what else I could try to bring the Rw/Rf down by 1-2% if possible, just wondering if there are any other strategies I could try.

Thank you,

George

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

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

George, it's difficult to tell without looking at models before and after modifications, and data. Did refinement converge (do you need to run more cycles)? Are weights optimal? etc etc... Pavel On 1/12/15 1:10 PM, George Devaniranjan wrote:

I have a related question to this.

I have a PDB structure (not mine I downloaded it from RCSB) and wondered if I could improve some of the problems in it using Phenix.

The Rw/R-free for the downloaded structure was 0.1279/0.1390

After I improved (or so I think so) the Rw/R-free is 0.1326/0.1440

The structure I have seems better in many aspects including its fit to the map (COOT/RSCC) + RMS Angles

I am a little confused whether to accept it or think I might have to do more validation checks as something has gone wrong?

Thank you, George

On Mon, Jan 12, 2015 at 4:01 PM, Pavel Afonine mailto:[email protected]> wrote:

Hi Tim,

it is a shame that I can't remember that particular paper where I think Randy Read or Axel Brunger have demonstrated that phase error and Rfree are well correlated. So if you get best Rwork and Rfree chances are you also get best map too.

All the best, Pavel

On 1/12/15 3:54 AM, Tim Gruene wrote:

...

Dear George,

what are the reasons why are you trying to reduce the R/Rfree instead of producing the best possible map?

Best, Tim

On 01/11/2015 11:24 PM, George D wrote:

...

Hi Phenix users,

I was looking at the phenix-refine manual and it mentions running

optimize_mask=true

optimize_xyz_weight=true

optimize_adp_weight=true

I was wondering what else I could try to bring the Rw/Rf down by 1-2% if possible, just wondering if there are any other strategies I could try.

Thank you,

George

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

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

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

Hi, I am sorry to ask another question, this is related to the above hence following up on this thread. I can calculate R/R-free using the validation tool but how does one calculate the R factors for the highest resolution shell ? Thank you, George On Tue, Jan 13, 2015 at 2:31 AM, Tim Gruene wrote:

Hi Pavel,

I was puzzled and misled by the choice of terms and implied an intent that apparently wasn't there.

Best, Tim

On 01/12/2015 10:01 PM, Pavel Afonine wrote:

...

Hi Tim,

it is a shame that I can't remember that particular paper where I think Randy Read or Axel Brunger have demonstrated that phase error and Rfree are well correlated. So if you get best Rwork and Rfree chances are you also get best map too.

All the best, Pavel

On 1/12/15 3:54 AM, Tim Gruene wrote:

...

Dear George,

what are the reasons why are you trying to reduce the R/Rfree instead of producing the best possible map?

Best, Tim

On 01/11/2015 11:24 PM, George D wrote:

...

Hi Phenix users,

I was looking at the phenix-refine manual and it mentions running

optimize_mask=true

optimize_xyz_weight=true

optimize_adp_weight=true

I was wondering what else I could try to bring the Rw/Rf down by 1-2% if possible, just wondering if there are any other strategies I could try.

Thank you,

George

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

_______________________________________________ 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

Hi Pavel, Thank you. If I wanted to compare a RCSB structure's highest resolution shell R with my refined highest resolution shell R. Do I generate a log file with the binned R *for the RCSB structure* by running phenix.refine but with *strategy='none' * or is there a way to generate these binned R factors for the deposited structure without running a refine cycle with none as strategy? Thank you, George On Tue, Jan 13, 2015 at 11:04 AM, Pavel Afonine wrote:

Hi George,

look in phenix.refine log file: it prints statistics in resolution bins (R-factors, completeness, number of reflections, <F>, etc); so you can take the number from highest resolution shell. Of course the definition of "highest resolution shell" is somewhat arbitrary, and also depends on how you do binning.

Pavel

On 1/13/15 6:48 AM, George Devaniranjan wrote:

Hi,

I am sorry to ask another question, this is related to the above hence following up on this thread.

I can calculate R/R-free using the validation tool but how does one calculate the

R factors for the highest resolution shell ?

Thank you, George

On Tue, Jan 13, 2015 at 2:31 AM, Tim Gruene wrote:

...

Hi Pavel,

I was puzzled and misled by the choice of terms and implied an intent that apparently wasn't there.

Best, Tim

On 01/12/2015 10:01 PM, Pavel Afonine wrote:

...

Hi Tim,

it is a shame that I can't remember that particular paper where I think Randy Read or Axel Brunger have demonstrated that phase error and Rfree are well correlated. So if you get best Rwork and Rfree chances are you also get best map too.

All the best, Pavel

On 1/12/15 3:54 AM, Tim Gruene wrote:

...

Dear George,

what are the reasons why are you trying to reduce the R/Rfree instead of producing the best possible map?

Best, Tim

On 01/11/2015 11:24 PM, George D wrote:

...

Hi Phenix users,

I was looking at the phenix-refine manual and it mentions running

optimize_mask=true

optimize_xyz_weight=true

optimize_adp_weight=true

I was wondering what else I could try to bring the Rw/Rf down by 1-2% if possible, just wondering if there are any other strategies I could try.

Thank you,

George

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

_______________________________________________ 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

Hi George, phenix.refine outputs this statistics at the very beginning of refinement and at the end, so both pieces of information that you need are in one log file. Pavel On 1/13/15 8:24 AM, George Devaniranjan wrote:

Hi Pavel, Thank you.

If I wanted to compare a RCSB structure's highest resolution shell R with my refined highest resolution shell R.

Do I generate a log file with the binned R _*for the RCSB structure*_ by running phenix.refine but with *_strategy='none' _*

or is there a way to generate these binned R factors for the deposited structure without running a refine cycle with none as strategy?

Thank you, George

On Tue, Jan 13, 2015 at 11:04 AM, Pavel Afonine mailto:[email protected]> wrote:

Hi George,

look in phenix.refine log file: it prints statistics in resolution bins (R-factors, completeness, number of reflections, <F>, etc); so you can take the number from highest resolution shell. Of course the definition of "highest resolution shell" is somewhat arbitrary, and also depends on how you do binning.

Pavel

On 1/13/15 6:48 AM, George Devaniranjan wrote:

...

Hi,

I am sorry to ask another question, this is related to the above hence following up on this thread.

I can calculate R/R-free using the validation tool but how does one calculate the

R factors for the highest resolution shell ?

Thank you, George

On Tue, Jan 13, 2015 at 2:31 AM, Tim Gruene mailto:[email protected]> wrote:

Hi Pavel,

I was puzzled and misled by the choice of terms and implied an intent that apparently wasn't there.

Best, Tim

On 01/12/2015 10:01 PM, Pavel Afonine wrote: > Hi Tim, > > it is a shame that I can't remember that particular paper where I think > Randy Read or Axel Brunger have demonstrated that phase error and Rfree > are well correlated. So if you get best Rwork and Rfree chances are you > also get best map too. > > All the best, > Pavel > > On 1/12/15 3:54 AM, Tim Gruene wrote: >> Dear George, >> >> what are the reasons why are you trying to reduce the R/Rfree instead of >> producing the best possible map? >> >> Best, >> Tim >> >> On 01/11/2015 11:24 PM, George D wrote: >>> Hi Phenix users, >>> >>> I was looking at the phenix-refine manual and it mentions running >>> >>> optimize_mask=true >>> >>> optimize_xyz_weight=true >>> >>> optimize_adp_weight=true >>> >>> >>> I was wondering what else I could try to bring the Rw/Rf down by 1-2% >>> if possible, just wondering if there are any other strategies I could >>> try. >>> >>> >>> Thank you, >>> >>> George >>> >>> >>> >>> _______________________________________________ >>> phenixbb mailing list >>> [email protected] mailto:[email protected] >>> http://phenix-online.org/mailman/listinfo/phenixbb >>> >> >> >> _______________________________________________ >> phenixbb mailing list >> [email protected] mailto:[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] mailto:[email protected] http://phenix-online.org/mailman/listinfo/phenixbb