Hi Tim, If you are using Phenix 1.5.2, that would not have had the new bulk-solvent model that Pavel is talking about. Are you using that old of a version for your tests? Jeff On Tue, Apr 17, 2012 at 8:07 AM, Pavel Afonine wrote:

Hi Tim,

this is because recently (a few weeks ago) we introduced a novel bulk-solvent model and a new overall anisotropic scaling procedure.

Some detailed overview is here:

http://www.phenix-online.org/presentations/PhenixSantaFe2012_PA.pdf

This is going to be published soon.

Pavel

On 4/17/12 7:59 AM, Timothy Springer wrote:

In one of the structures I have refined, I found phenix 1.5_2 gave 1 to 1.5% lower Rfree and Rwork than REFMAC 5.5.0102. Can I propose in a publication that "We attribute these differences to methods for handling bulk solvent in PHENIX that are less adversely affected by deficiencies in the crystallographic data."? This is a guess, because there is a large contribution to refinement from bulk solvent, rather than factual information. I did cross-refinement, and found it going from structures refined in REfmac to Phenix, and vice versa. The Phenix maps looked better, too. Is this a reasonable suggestion, and would there be a relevant reference? Tim

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

Hi Tim, How can you say the difference is really down to bulk solvent correction? there's several other parameters that affect differences in R factors.... Just curious R On 17 Apr 2012, at 16:27, Timothy Springer wrote: no, it just takes me two years to write up sometimes--- So this is a historical question. But there is a ref in that powerpoint to following ref. If this was implemented in Phenix and not Refmac 2 years ago, I should cite it. Acta Crystallogr D Biol Crystallogr.http://www.ncbi.nlm.nih.gov.ezp-prod1.hul.harvard.edu/pubmed?term=robust%20b... 2005 Jul;61(Pt 7):850-5. Epub 2005 Jun 24. A robust bulk-solvent correction and anisotropic scaling procedure. Afonine PVhttp://www.ncbi.nlm.nih.gov.ezp-prod1.hul.harvard.edu/pubmed?term=%22Afonine..., Grosse-Kunstleve RWhttp://www.ncbi.nlm.nih.gov.ezp-prod1.hul.harvard.edu/pubmed?term=%22Grosse-..., Adams PDhttp://www.ncbi.nlm.nih.gov.ezp-prod1.hul.harvard.edu/pubmed?term=%22Adams%2.... Source Lawrence Berkeley National Laboratory, One Cyclotron Road, Building 64R0121, Berkeley, CA 94720, USA. [email protected]mailto:[email protected] Tim On Apr 17, 2012, at 11:18 AM, Jeff Headd wrote: Hi Tim, If you are using Phenix 1.5.2, that would not have had the new bulk-solvent model that Pavel is talking about. Are you using that old of a version for your tests? Jeff On Tue, Apr 17, 2012 at 8:07 AM, Pavel Afonine mailto:[email protected]> wrote: Hi Tim, this is because recently (a few weeks ago) we introduced a novel bulk-solvent model and a new overall anisotropic scaling procedure. Some detailed overview is here: http://www.phenix-online.org/presentations/PhenixSantaFe2012_PA.pdf This is going to be published soon. Pavel On 4/17/12 7:59 AM, Timothy Springer wrote: In one of the structures I have refined, I found phenix 1.5_2 gave 1 to 1.5% lower Rfree and Rwork than REFMAC 5.5.0102. Can I propose in a publication that "We attribute these differences to methods for handling bulk solvent in PHENIX that are less adversely affected by deficiencies in the crystallographic data."? This is a guess, because there is a large contribution to refinement from bulk solvent, rather than factual information. I did cross-refinement, and found it going from structures refined in REfmac to Phenix, and vice versa. The Phenix maps looked better, too. Is this a reasonable suggestion, and would there be a relevant reference? Tim _______________________________________________ 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 Roberto Steiner, PhD Group Leader Randall Division of Cell and Molecular Biophysics King's College London Room 3.10A New Hunt's House Guy's Campus SE1 1UL, London, UK Tel 0044-20-78488216 Fax 0044-20-78486435 [email protected]mailto:[email protected]

difference density on the backbone can also arise (more often) due to suboptimal ADP refinement/restraints..or a bug in a particular build.. best wishes roberto PS I cc-ed this email to Garib in case he's got some info specific for 5.5.0102 On 17 Apr 2012, at 17:22, Timothy Springer wrote: As I said, I cant say that, it is a guess based on two facts. 1) Looking at the output of refinement steps, and seeing that the largest decrease in R during refinement comes in the bulk solvent correction step. 2) It is also based on the common occurrence in the Refmac but not Phenix maps of negative difference density on backbone. This is a symptom of errors in bulk solvent correction. Tim On Apr 17, 2012, at 12:06 PM, Steiner, Roberto wrote: Hi Tim, How can you say the difference is really down to bulk solvent correction? there's several other parameters that affect differences in R factors.... Just curious R On 17 Apr 2012, at 16:27, Timothy Springer wrote: no, it just takes me two years to write up sometimes--- So this is a historical question. But there is a ref in that powerpoint to following ref. If this was implemented in Phenix and not Refmac 2 years ago, I should cite it. Acta Crystallogr D Biol Crystallogr.http://www.ncbi.nlm.nih.gov.ezp-prod1.hul.harvard.edu/pubmed?term=robust%20b... 2005 Jul;61(Pt 7):850-5. Epub 2005 Jun 24. A robust bulk-solvent correction and anisotropic scaling procedure. Afonine PVhttp://www.ncbi.nlm.nih.gov.ezp-prod1.hul.harvard.edu/pubmed?term=%22Afonine..., Grosse-Kunstleve RWhttp://www.ncbi.nlm.nih.gov.ezp-prod1.hul.harvard.edu/pubmed?term=%22Grosse-..., Adams PDhttp://www.ncbi.nlm.nih.gov.ezp-prod1.hul.harvard.edu/pubmed?term=%22Adams%2.... Source Lawrence Berkeley National Laboratory, One Cyclotron Road, Building 64R0121, Berkeley, CA 94720, USA. [email protected]mailto:[email protected] Tim On Apr 17, 2012, at 11:18 AM, Jeff Headd wrote: Hi Tim, If you are using Phenix 1.5.2, that would not have had the new bulk-solvent model that Pavel is talking about. Are you using that old of a version for your tests? Jeff On Tue, Apr 17, 2012 at 8:07 AM, Pavel Afonine mailto:[email protected]> wrote: Hi Tim, this is because recently (a few weeks ago) we introduced a novel bulk-solvent model and a new overall anisotropic scaling procedure. Some detailed overview is here: http://www.phenix-online.org/presentations/PhenixSantaFe2012_PA.pdf This is going to be published soon. Pavel On 4/17/12 7:59 AM, Timothy Springer wrote: In one of the structures I have refined, I found phenix 1.5_2 gave 1 to 1.5% lower Rfree and Rwork than REFMAC 5.5.0102. Can I propose in a publication that "We attribute these differences to methods for handling bulk solvent in PHENIX that are less adversely affected by deficiencies in the crystallographic data."? This is a guess, because there is a large contribution to refinement from bulk solvent, rather than factual information. I did cross-refinement, and found it going from structures refined in REfmac to Phenix, and vice versa. The Phenix maps looked better, too. Is this a reasonable suggestion, and would there be a relevant reference? Tim _______________________________________________ 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 Roberto Steiner, PhD Group Leader Randall Division of Cell and Molecular Biophysics King's College London Room 3.10A New Hunt's House Guy's Campus SE1 1UL, London, UK Tel 0044-20-78488216 Fax 0044-20-78486435 [email protected]mailto:[email protected] _______________________________________________ phenixbb mailing list [email protected]mailto:[email protected] http://phenix-online.org/mailman/listinfo/phenixbb Roberto Steiner, PhD Group Leader Randall Division of Cell and Molecular Biophysics King's College London Room 3.10A New Hunt's House Guy's Campus SE1 1UL, London, UK Tel 0044-20-78488216 Fax 0044-20-78486435 [email protected]mailto:[email protected]