Re: [ccp4bb] TLS refinement refmac
Well - most refinement procedures output the Fobs and the Fcalc on the same (more-or-less) absolute scale.. After data processing the observations are on a more or less arbitrary scale, so that seems sensible. It gets more problematic when you start correcting for anisotropy.. Eleanor On 17 July 2013 15:59, Phoebe A. Rice pr...@uchicago.edu wrote: EEK, it seems to me that anything called simply FP should be unadulterated FP. If the software modifies a column in some way, it should give it a new label, shouldn't it? ++ Phoebe A. Rice Dept. of Biochemistry Molecular Biology The University of Chicago 773 834 1723; pr...@uchicago.edu http://bmb.bsd.uchicago.edu/Faculty_and_Research/ http://www.rsc.org/shop/books/2008/9780854042722.asp -- *From:* CCP4 bulletin board [CCP4BB@JISCMAIL.AC.UK] on behalf of Eleanor Dodson [eleanor.dod...@york.ac.uk] *Sent:* Wednesday, July 17, 2013 6:05 AM *To:* CCP4BB@JISCMAIL.AC.UK *Subject:* Re: [ccp4bb] TLS refinement refmac Oh dear - you should always start refinement against the original processed data - but others have told you that already.. The TLS files will not be appropriate for the rescaled FPs - it is surprising that the Rfactor actually goes down though! Or are you doing several cycles of refinement in thesecond pass too - in that case one would hope the Rfactor would continue to fall. Eleanor On 17 July 2013 11:00, Guenter Fritz guenter.fr...@uni-konstanz.dewrote: Am 17.07.2013 11:59, schrieb Guenter Fritz: Hi Stefan and Gottfried, thanks a lot for the answers. This is the point. Wouldn't it make more sense to add an extra column that contains the changed Fs? Best, Guenter Hi, it is strongly advised to use the original mtz e.g. scala.mtz as the refmac input mtz in all refmac runs, as this contains the original Fs - Refmac applies some aniso corrections to the Fs and puts them into the output.mtz. so the output Fs are not the same as in the input F - therefore one should use the scala.mtz cheers Stefan -Ursprüngliche Nachricht- Von: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] Im Auftrag von Guenter Fritz Gesendet: Mittwoch, 17. Juli 2013 10:39 An: CCP4BB@JISCMAIL.AC.UK Betreff: [ccp4bb] TLS refinement refmac Dear all, one gets different R values, if you re-read in the mtz written out by refmac after TLS refinement. I think this issue had been a while ago in ccp4bb, but I can't find the right track. Here are the details. 1st run: If we do TLS + restr. refinement in refmac we get: InitialFinal R factor0.3010 0.2170 R free0.3175 0.2695 2nd run: Now, if we use the same input pdb and the same input tls paramter file, but use the mtz written out by the first run, we get: InitialFinal R factor0.2274 0.1903 R free0.2482 0.2540 Apparently in the mtz file written out by the 1st refmac must contain some information that is re-read in the 2nd run. But one just defines FP, SIGFP and Rfree flags. Do FPs change in the output mtz after TLS refinement?? Any help to clarify this is appreciated. Thanks, Guenter
Re: [ccp4bb] TLS refinement refmac
Am 17.07.2013 11:59, schrieb Guenter Fritz: Hi Stefan and Gottfried, thanks a lot for the answers. This is the point. Wouldn't it make more sense to add an extra column that contains the changed Fs? Best, Guenter Hi, it is strongly advised to use the original mtz e.g. scala.mtz as the refmac input mtz in all refmac runs, as this contains the original Fs - Refmac applies some aniso corrections to the Fs and puts them into the output.mtz. so the output Fs are not the same as in the input F - therefore one should use the scala.mtz cheers Stefan -Ursprüngliche Nachricht- Von: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] Im Auftrag von Guenter Fritz Gesendet: Mittwoch, 17. Juli 2013 10:39 An: CCP4BB@JISCMAIL.AC.UK Betreff: [ccp4bb] TLS refinement refmac Dear all, one gets different R values, if you re-read in the mtz written out by refmac after TLS refinement. I think this issue had been a while ago in ccp4bb, but I can't find the right track. Here are the details. 1st run: If we do TLS + restr. refinement in refmac we get: InitialFinal R factor0.3010 0.2170 R free0.3175 0.2695 2nd run: Now, if we use the same input pdb and the same input tls paramter file, but use the mtz written out by the first run, we get: InitialFinal R factor0.2274 0.1903 R free0.2482 0.2540 Apparently in the mtz file written out by the 1st refmac must contain some information that is re-read in the 2nd run. But one just defines FP, SIGFP and Rfree flags. Do FPs change in the output mtz after TLS refinement?? Any help to clarify this is appreciated. Thanks, Guenter
Re: [ccp4bb] TLS refinement refmac
Oh dear - you should always start refinement against the original processed data - but others have told you that already.. The TLS files will not be appropriate for the rescaled FPs - it is surprising that the Rfactor actually goes down though! Or are you doing several cycles of refinement in thesecond pass too - in that case one would hope the Rfactor would continue to fall. Eleanor On 17 July 2013 11:00, Guenter Fritz guenter.fr...@uni-konstanz.de wrote: Am 17.07.2013 11:59, schrieb Guenter Fritz: Hi Stefan and Gottfried, thanks a lot for the answers. This is the point. Wouldn't it make more sense to add an extra column that contains the changed Fs? Best, Guenter Hi, it is strongly advised to use the original mtz e.g. scala.mtz as the refmac input mtz in all refmac runs, as this contains the original Fs - Refmac applies some aniso corrections to the Fs and puts them into the output.mtz. so the output Fs are not the same as in the input F - therefore one should use the scala.mtz cheers Stefan -Ursprüngliche Nachricht- Von: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] Im Auftrag von Guenter Fritz Gesendet: Mittwoch, 17. Juli 2013 10:39 An: CCP4BB@JISCMAIL.AC.UK Betreff: [ccp4bb] TLS refinement refmac Dear all, one gets different R values, if you re-read in the mtz written out by refmac after TLS refinement. I think this issue had been a while ago in ccp4bb, but I can't find the right track. Here are the details. 1st run: If we do TLS + restr. refinement in refmac we get: InitialFinal R factor0.3010 0.2170 R free0.3175 0.2695 2nd run: Now, if we use the same input pdb and the same input tls paramter file, but use the mtz written out by the first run, we get: InitialFinal R factor0.2274 0.1903 R free0.2482 0.2540 Apparently in the mtz file written out by the 1st refmac must contain some information that is re-read in the 2nd run. But one just defines FP, SIGFP and Rfree flags. Do FPs change in the output mtz after TLS refinement?? Any help to clarify this is appreciated. Thanks, Guenter
Re: [ccp4bb] TLS refinement refmac
EEK, it seems to me that anything called simply FP should be unadulterated FP. If the software modifies a column in some way, it should give it a new label, shouldn't it? ++ Phoebe A. Rice Dept. of Biochemistry Molecular Biology The University of Chicago 773 834 1723; pr...@uchicago.edumailto:pr...@uchicago.edu http://bmb.bsd.uchicago.edu/Faculty_and_Research/ http://www.rsc.org/shop/books/2008/9780854042722.asp From: CCP4 bulletin board [CCP4BB@JISCMAIL.AC.UK] on behalf of Eleanor Dodson [eleanor.dod...@york.ac.uk] Sent: Wednesday, July 17, 2013 6:05 AM To: CCP4BB@JISCMAIL.AC.UK Subject: Re: [ccp4bb] TLS refinement refmac Oh dear - you should always start refinement against the original processed data - but others have told you that already.. The TLS files will not be appropriate for the rescaled FPs - it is surprising that the Rfactor actually goes down though! Or are you doing several cycles of refinement in thesecond pass too - in that case one would hope the Rfactor would continue to fall. Eleanor On 17 July 2013 11:00, Guenter Fritz guenter.fr...@uni-konstanz.demailto:guenter.fr...@uni-konstanz.de wrote: Am 17.07.2013 11:59, schrieb Guenter Fritz: Hi Stefan and Gottfried, thanks a lot for the answers. This is the point. Wouldn't it make more sense to add an extra column that contains the changed Fs? Best, Guenter Hi, it is strongly advised to use the original mtz e.g. scala.mtz as the refmac input mtz in all refmac runs, as this contains the original Fs - Refmac applies some aniso corrections to the Fs and puts them into the output.mtz. so the output Fs are not the same as in the input F - therefore one should use the scala.mtz cheers Stefan -Ursprüngliche Nachricht- Von: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UKmailto:CCP4BB@JISCMAIL.AC.UK] Im Auftrag von Guenter Fritz Gesendet: Mittwoch, 17. Juli 2013 10:39 An: CCP4BB@JISCMAIL.AC.UKmailto:CCP4BB@JISCMAIL.AC.UK Betreff: [ccp4bb] TLS refinement refmac Dear all, one gets different R values, if you re-read in the mtz written out by refmac after TLS refinement. I think this issue had been a while ago in ccp4bb, but I can't find the right track. Here are the details. 1st run: If we do TLS + restr. refinement in refmac we get: InitialFinal R factor0.3010 0.2170 R free0.3175 0.2695 2nd run: Now, if we use the same input pdb and the same input tls paramter file, but use the mtz written out by the first run, we get: InitialFinal R factor0.2274 0.1903 R free0.2482 0.2540 Apparently in the mtz file written out by the 1st refmac must contain some information that is re-read in the 2nd run. But one just defines FP, SIGFP and Rfree flags. Do FPs change in the output mtz after TLS refinement?? Any help to clarify this is appreciated. Thanks, Guenter