Re: [ccp4bb] molecule on symmetry axis
Dear colleagues, Thank you for your suggestions. I was able to do the refinement of the sulfates and tartaric acid in phenix by changing the occupancies to 0.33 and 0.5, It appears to me (from the log file and from graphics afterwards) that the program automatically turns off the interaction with the symmetry mate when one puts the right occupancy.. Jackie Vitali
Re: [ccp4bb] refining structures with engineered disulfide bonds
I think REFMAC will generate anappropriate SSBOND entry in the pdb for you if you run it from the GUI with Review restraintsoption Eleanor On 10/20/2010 08:40 PM, Seema Mittal wrote: Hi All, I have engineered intra-molecular disulfide bond in my protein monomer. The protein functions as a homodimer. In crystal structure, there is clean electron density for the S-S bond in one monomer (bond length 2.2A), but there seems to be slightly messy density for the same in the other monomer with (S-S bond length of 2.7A). An alternate conformation of one of the cys seems plausible on the messy side. There is considerable negative density associated with this region in both monomers, more so on the messy side. My question is : do i need to select additional parameters or define any sort of constraints during refinement, in order to refine this introduced covalent bond? Thanks much for your help. Best Seema Mittal Graduate Research Assistant Department of Biochemistry Molecular Pharmacology, 970L Lazare Research Building, University of Massachusetts Medical School, 364 Plantation Street, Worcester, MA 01605
Re: [ccp4bb] FOM: Phaser vs SigmaA
Being lazy I would just do refinement with REFMAC and let it generate the SigmaA values.. then try rebuilding with either buccaneer or Arp/Warp.They willboth generate weighting terms based on the Rfactors. Eleanor On 10/21/2010 12:58 AM, Goragot Wisedchaisri wrote: Hi, I have helices that I did rigid body refinement with Phaser (after phased rotation and phased translation in Molrep). I compare FOM output by Phaser to the FOM computed by sigmaA using the Phaser refined coordinates and found that FOM from Phaser is only about half (~0.25) of FOM from SigmaA (~0.5). I'm running Phaser using ccp4 version 6.1.13. I remember a while back that Phaser used to calculate a priori sigmaA estimation based on assumed model rmsd error. I am not sure if this a priori SigmaA weight is also output in the FOM column. If this is not the case, could anyone point me to a documentation of how Phaser calculates FOM. The Phaser wiki and J App Crys paper does not seem to have detail information on this. I could just use SigmaA or do refinement in Refmac but I have to say that I like the low FOM from Phaser because model bias seem to be much less after density modification. It also saves me from having to blur the phase probability distribution in order to down weight FOM when FOM is too high. But I still would like to know how Phaser currently calculates the unusually low output FOM. Many thanks, George Wisedchaisri
[ccp4bb] Biological Structures Group BCA Winter Meeting - Wed 15th December, University of Reading - REGISTRATION NOW OPEN!
Dear All, At last, registration and details of the BSG Winter Meeting can be found at: http://www.reading.ac.uk/biologicalsciences/businessdevelopment/biosci-B CAwintermeet.aspx We look forward to welcoming you to Reading! Yours, Kim Watson
[ccp4bb] Regarding space group P1, P21
Dear CCP4BB members, I have solved a protein-drug complex structure in P21212 space group. In this structure, the drug molecule is falling on the two-fold symmetry axis having averaged electron density with 0.5 occupancy. We tried a lot to crystallize this protein-drug complex in different space group but no success so far. I have tried to solve the same data in space group P1 (statistics are fine as I have collected data for 360 degree). The map looks even better with one conformation for a drug. Interestingly, then I reprocessed the same data using imosflm in P21 space group which have penalty 1 compared to 4 for P21212. The structure in P21 is also refining well (with one conformation of the drug compound without symmetry axis at the ligand position). The question is , is it a good practice to solve this structure in P1 and P21 even if the data has higher symmetry? Secondly, I have been advised that I have to be careful to refine structure in P1 as there will be problem regarding observation/parameter ratio if I add too many water molecules. What will be the case if the electron density present for water molecules? I can put restrains to protein structure but I am just curious to know one restrain equals how many observations. I look forward to hear your suggestions. Kind regards, Mohinder Pal
Re: [ccp4bb] Regarding space group P1, P21
Hi Since you're using iMosflm to process the data, it is well worthwhile running the Quickscale task following integration (I would actually run it after integrating ~5 - 10 degrees of data) to see if the true crystal symmetry determined by analysing agreement of the intensities of symmetry related reflections is actually the same as that indicated by the penalties from indexing. Remember that the relationship between the unit cell dimensions is a consequence of the true symmetry, not vice versa - most crystallographers who have been in the game more than a few years have examples of lower symmetry crystals with apparently higher symmetry cell dimensions - a relatively common occurrence to have cell dimensions that look right for tetragonal when the true symmetry is orthorhombic. Of course, following integration scaling etc you would probably want to check for things like twinning etc... In general, I think you should probably solve and refine in the highest symmetry space group that is most consistent with your data. If the experiment gives you just as good results in the higher symmetry space group as the lower, I would go for the higher symmetry. In your case, if P21 solution/refinement is as good as P1, but both are better than P21212, I would tend towards using the P21 solution. On 21 Oct 2010, at 12:05, Mohinder Pal wrote: Dear CCP4BB members, I have solved a protein-drug complex structure in P21212 space group. In this structure, the drug molecule is falling on the two-fold symmetry axis having averaged electron density with 0.5 occupancy. We tried a lot to crystallize this protein-drug complex in different space group but no success so far. I have tried to solve the same data in space group P1 (statistics are fine as I have collected data for 360 degree). The map looks even better with one conformation for a drug. Interestingly, then I reprocessed the same data using imosflm in P21 space group which have penalty 1 compared to 4 for P21212. The structure in P21 is also refining well (with one conformation of the drug compound without symmetry axis at the ligand position). The question is , is it a good practice to solve this structure in P1 and P21 even if the data has higher symmetry? Secondly, I have been advised that I have to be careful to refine structure in P1 as there will be problem regarding observation/parameter ratio if I add too many water molecules. What will be the case if the electron density present for water molecules? I can put restrains to protein structure but I am just curious to know one restrain equals how many observations. I look forward to hear your suggestions. Kind regards, Mohinder Pal Harry -- Dr Harry Powell, MRC Laboratory of Molecular Biology, MRC Centre, Hills Road, Cambridge, CB2 0QH
[ccp4bb] tlsanl and phenix TLS results
[ ccp4 6.1.3 ] i have some phenix TLS tensors i'd like to evaluate in tlsanl [*]. are there specific conversions/transformations to be wary of when setting up the job or interpreting the output? -bryan [*] originally posted on phenix BB with some gory details.
Re: [ccp4bb] tlsanl and phenix TLS results
Hi Bryan That would obviously depend on the precise definition and order of the T, L S tensor values that phenix writes out; not being a phenix user I can't help you there, but I assume that the phenix documentation will tell you all you need to know. The format required by TLSANL is defined in the REFMAC5 documentation here: http://www.ccp4.ac.uk/dist/html/refmac5/files/tls.html Note that the first line starting with '!' and containing the word 'REFMAC' may look like a comment but is actually critical, since it tells TLSANL that the data is in REFMAC format. There's also an option BRESID in TLSANL which you will need to set if the atomic Biso values in the file are 'residual' values (i.e. not including the TLS contribution); again the phenix documentation will no doubt enlighten you on this point. It may well be that phenix writes out the values in a format which is already compatible with REFMAC TLSANL. That I don't know, but if it does it makes it very easy, since you won't have to change anything. Others have already pointed out that it would be better if TLSANL read the values in the PDB header (assuming of course that they are present!). This is on my to-do list. Cheers -- Ian On Thu, Oct 21, 2010 at 1:26 PM, Bryan Lepore bryanlep...@gmail.com wrote: [ ccp4 6.1.3 ] i have some phenix TLS tensors i'd like to evaluate in tlsanl [*]. are there specific conversions/transformations to be wary of when setting up the job or interpreting the output? -bryan [*] originally posted on phenix BB with some gory details.
Re: [ccp4bb] tlsanl *NOW ALSO REFMAC* and phenix TLS results
documentation then i conclude the TLS protocol in refmac is markedly different from phenix (i know this is not strictly a ccp4 question). cf. : refmac: TLS RANGE 'A 245.' 'A 252.' ALL ORIGIN14.019 -10.476 -35.068 T 0.4974 0.0372 0.3453 0.0674 0.2984 0.0431 L 21.5463 29.5753 20.7545 39.3304 9.2958 -11.5975 S 0.6432 1.0787 -0.1449 1.2672 2.1065 1.5489 -1.1854 -0.0349 center of reaction : 18.652 -5.646 -40.945 from orth. axes libration perpendicular to TLS plane : 13.256 deg^2 phenix: TLS RANGE 'A 245.' 'A 252.' ALL ORIGIN 14.019 -10.476 -35.068 T 0.2961 0.2533 0.2969 -0.0097 0.0111 0.0205 L 0.0006 0.0005 0.0006 0.0006 -0.0002 -0.0001 S 0.0058 -0.0026 -0.0027 0.0035 -0.0017 0.0011 -0.0052 0.0027 center of reaction : -606.636-734.593 -54.324 from orth. axes libration perpendicular to TLS plane : 0.001 deg^2 (trying to keep this a refmac/tlsanl question) and the results are equivalent - because of factors perhaps in symmetrization of S, because there are no unique choices of these things. if so, only the refmac axes faciliate an intuitive comparison. regards, -bryan p.s: link to original thread phenix board : http://www.phenix-online.org/pipermail/phenixbb/2010-October/004724.html
Re: [ccp4bb] Regarding space group P1, P21
There is nothing fundamentally wrong with refining in P1 even if the P21212 symmetry is present. An effective way to reduce the number of parameters wold be to introduce tight restraints. If you decide to lower the symmetry, go with P21 as it still keeps your ligand off symmetry axes. You can then add tight ncs restraints for the protein part. Alternatively, you can finish up the refinement in P21212 but get the maps for your publication drawn in P21 (with appropriate explanation). The reason to use the highest symmetry possible is because it presumably gives you a more precise structure since data quality may be better in P21212. I am not quite sure what you mean by putting restraints on protein - NCS? If so, tight restraints should approximately reduce the number of effective parameters by the number of copies. It appears (perhaps someone will correct me) that *constraints* are only available in CNS, but tight restraints supposedly approach that limit. Ed. On Thu, 2010-10-21 at 13:05 +0100, Mohinder Pal wrote: Dear CCP4BB members, I have solved a protein-drug complex structure in P21212 space group. In this structure, the drug molecule is falling on the two-fold symmetry axis having averaged electron density with 0.5 occupancy. We tried a lot to crystallize this protein-drug complex in different space group but no success so far. I have tried to solve the same data in space group P1 (statistics are fine as I have collected data for 360 degree). The map looks even better with one conformation for a drug. Interestingly, then I reprocessed the same data using imosflm in P21 space group which have penalty 1 compared to 4 for P21212. The structure in P21 is also refining well (with one conformation of the drug compound without symmetry axis at the ligand position). The question is , is it a good practice to solve this structure in P1 and P21 even if the data has higher symmetry? Secondly, I have been advised that I have to be careful to refine structure in P1 as there will be problem regarding observation/parameter ratio if I add too many water molecules. What will be the case if the electron density present for water molecules? I can put restrains to protein structure but I am just curious to know one restrain equals how many observations. I look forward to hear your suggestions. Kind regards, Mohinder Pal -- I'd jump in myself, if I weren't so good at whistling. Julian, King of Lemurs
Re: [ccp4bb] Regarding space group P1, P21
Dear Mohinder and Ed, If you process your data in a lower symmetry space group, you will have more unique reflections, since reflections which are related by the higher symmetry will be avaraged during scaling in a higher symmetry space group (i.e. a 2fold or 3fold axis), while in lower symmetry space groups they will not. So the observation to parameter ratio stays the same and is only depending on resolution and solvent content. The question one has to ask of course is: are these reflections really different, or are they the same only not averaged? In the latter case, you have more reflections, but not more information. As Ed mentions, using tight NCS restraints would in this case mimick the crystallographic symmetry. I would calculate maps while leaving out the inhibitor (omit maps) and check that the inhibitor indeed has a unique conformation in the lower symmetry space group. In that case the symmetry of the inhibitor, and therefore of your crystal, is the lower symmetry. If the inhibitor has a twofold disorder in the lower symmetry space group, you really have a higher symmetry space group and should work with this space group. In that case you can fit a molecule on the twofold axis with an occupancy of 0.5 and Refmac will automatically recognize the special position. Best regards, Herman -Original Message- From: CCP4 bulletin board [mailto:ccp...@jiscmail.ac.uk] On Behalf Of Ed Pozharski Sent: Thursday, October 21, 2010 5:05 PM To: CCP4BB@JISCMAIL.AC.UK Subject: Re: [ccp4bb] Regarding space group P1, P21 There is nothing fundamentally wrong with refining in P1 even if the P21212 symmetry is present. An effective way to reduce the number of parameters wold be to introduce tight restraints. If you decide to lower the symmetry, go with P21 as it still keeps your ligand off symmetry axes. You can then add tight ncs restraints for the protein part. Alternatively, you can finish up the refinement in P21212 but get the maps for your publication drawn in P21 (with appropriate explanation). The reason to use the highest symmetry possible is because it presumably gives you a more precise structure since data quality may be better in P21212. I am not quite sure what you mean by putting restraints on protein - NCS? If so, tight restraints should approximately reduce the number of effective parameters by the number of copies. It appears (perhaps someone will correct me) that *constraints* are only available in CNS, but tight restraints supposedly approach that limit. Ed. On Thu, 2010-10-21 at 13:05 +0100, Mohinder Pal wrote: Dear CCP4BB members, I have solved a protein-drug complex structure in P21212 space group. In this structure, the drug molecule is falling on the two-fold symmetry axis having averaged electron density with 0.5 occupancy. We tried a lot to crystallize this protein-drug complex in different space group but no success so far. I have tried to solve the same data in space group P1 (statistics are fine as I have collected data for 360 degree). The map looks even better with one conformation for a drug. Interestingly, then I reprocessed the same data using imosflm in P21 space group which have penalty 1 compared to 4 for P21212. The structure in P21 is also refining well (with one conformation of the drug compound without symmetry axis at the ligand position). The question is , is it a good practice to solve this structure in P1 and P21 even if the data has higher symmetry? Secondly, I have been advised that I have to be careful to refine structure in P1 as there will be problem regarding observation/parameter ratio if I add too many water molecules. What will be the case if the electron density present for water molecules? I can put restrains to protein structure but I am just curious to know one restrain equals how many observations. I look forward to hear your suggestions. Kind regards, Mohinder Pal -- I'd jump in myself, if I weren't so good at whistling. Julian, King of Lemurs
Re: [ccp4bb] Regarding space group P1, P21
Dear Mohinder, On Thu, Oct 21, 2010 at 01:05:42PM +0100, Mohinder Pal wrote: The question is , is it a good practice to solve this structure in P1 and P21 even if the data has higher symmetry? On a slightly philosophical note regarding the final model (and not necessarily the 'good practice' leading to it): shouldn't our model describe the experiment (intensities from a crystal of given symmetry) and not the other way round (changing the experimental data to make modeling easier)? Or maybe I'm too strict here ... If your crystal has P21212 then I would model it this way: having a compound on a 2-fold with half occupancy isn't really a problem nowadays with modern refinement programs. And yes: it might confuse molecular biologists downloading the PDB file. And since their needs often dictate how we are supposed to produce models for our experiments, the time might come where all structures being refined in P1 with only the A chain deposited ;-) Cheers Clemens -- *** * Clemens Vonrhein, Ph.D. vonrhein AT GlobalPhasing DOT com * * Global Phasing Ltd. * Sheraton House, Castle Park * Cambridge CB3 0AX, UK *-- * BUSTER Development Group (http://www.globalphasing.com) ***
Re: [ccp4bb] Regarding space group P1, P21
Hi Herman, On Thu, Oct 21, 2010 at 05:31:51PM +0200, herman.schreu...@sanofi-aventis.com wrote: If you process your data in a lower symmetry space group, you will have more unique reflections, since reflections which are related by the higher symmetry will be avaraged during scaling in a higher symmetry space group (i.e. a 2fold or 3fold axis), while in lower symmetry space groups they will not. So the observation to parameter ratio stays the same and is only depending on resolution and solvent content. True - if you count Miller indices as observations. But if you think about information content than probably not (as you discuss below). The question one has to ask of course is: are these reflections really different, or are they the same only not averaged? Yes - by merging we're getting better data (better error estimate on the intensity due to higher multiplicity). So there isn't really independent information in 50% of the reflections if e.g. going from P21 to P1 - we've only increased the noise because the multiplicity of each reflection has been reduced. In the latter case, you have more reflections, but not more information. As Ed mentions, using tight NCS restraints would in this case mimick the crystallographic symmetry. Apart from the (good) NCS argument, one could go even further: We could also just collect 36000 degree of data on a 7A Lysozyme crystal and refine against completely unmerged data. After all, why should we stop at removing only the some symmetry operators from our data merging ... lets get rid of all of them including th x,y,z operator and use unmerged data. Then we could refine Lysozyme with anisotropic hydrogens and no restraints against 7A data since we have a huge number of 'observations' ... right? But seriously: there is a difference in having reflections (H, K, L) and independent data (I, SIGI). Maybe we should talk more about (independent observations)/parameters ratio in the same way we look at depdencies of parameters (e.g. restraints on Bfactors etc). Cheers Clemens -- *** * Clemens Vonrhein, Ph.D. vonrhein AT GlobalPhasing DOT com * * Global Phasing Ltd. * Sheraton House, Castle Park * Cambridge CB3 0AX, UK *-- * BUSTER Development Group (http://www.globalphasing.com) ***
Re: [ccp4bb] Regarding space group P1, P21
You pick the Rfree flags in the high-symmetry space group, and then use CAD with OUTLIM SPACE P1 to symmetry-expand them to P1 (or whatever you like). Things get trickier, however, when your NCS is close to, (bot not exactly) crystallographic (NECS?). Or if you are simply not sure. The best way I can think of to deal with this situation is to road test your Rfree: 1) do something that you know is wrong, like delete a helix, or put some side chains in the wrong place 2) refine with NCS turned on 3) check that Rfree actually goes up 4) un-do the wrong things 5) refine again 6) check that Rfree actually goes down 7) try again with NCS turned off Remembering these timeless words of wisdom: Control, Control, you must learn CONTROL! -Yoda (Jedi Master) -James Holton MAD Scientist On 10/21/2010 8:46 AM, Christina Bourne wrote: Dear all, How would one properly select reflections for R-free in these situations? Presumably if the selection is done in P1 then it mimics twinning or high NCS, such that reflections in both the work and free set will be (potentially?) related by symmetry. -Christina *From:* Mohinder Pal m...@soton.ac.uk *To:* CCP4BB@JISCMAIL.AC.UK *Sent:* Thu, October 21, 2010 7:05:42 AM *Subject:* [ccp4bb] Regarding space group P1, P21 Dear CCP4BB members, I have solved a protein-drug complex structure in P21212 space group. In this structure, the drug molecule is falling on the two-fold symmetry axis having averaged electron density with 0.5 occupancy. We tried a lot to crystallize this protein-drug complex in different space group but no success so far. I have tried to solve the same data in space group P1 (statistics are fine as I have collected data for 360 degree). The map looks even better with one conformation for a drug. Interestingly, then I reprocessed the same data using imosflm in P21 space group which have penalty 1 compared to 4 for P21212. The structure in P21 is also refining well (with one conformation of the drug compound without symmetry axis at the ligand position). The question is , is it a good practice to solve this structure in P1 and P21 even if the data has higher symmetry? Secondly, I have been advised that I have to be careful to refine structure in P1 as there will be problem regarding observation/parameter ratio if I add too many water molecules. What will be the case if the electron density present for water molecules? I can put restrains to protein structure but I am just curious to know one restrain equals how many observations. I look forward to hear your suggestions. Kind regards, Mohinder Pal
Re: [ccp4bb] Regarding space group P1, P21
I have heard many times that it is a black eye to refine in a lower-symmetry spacegroup, but I could never really understand why. The higher symmetry could be considered merely a helpful theoretical lens to improve signal-to-noise, and therefore imposing higher symmetry on the data could be seen as a sort of *leniency* of scientific (or at least empiric) rigor. I think similarly about using discrete spot intensities rather than the whole image--we assume Bragg conditions and neglect certain things about the image between the spots, which is usually valid, but not always. I wonder why it is considered maladroit to refine in a lower spacegroup, then--don't higher spacegroup impose more assumptions than p1? Jacob Keller - Original Message - From: James Holton To: CCP4BB@JISCMAIL.AC.UK Sent: Thursday, October 21, 2010 10:55 AM Subject: Re: [ccp4bb] Regarding space group P1, P21 You pick the Rfree flags in the high-symmetry space group, and then use CAD with OUTLIM SPACE P1 to symmetry-expand them to P1 (or whatever you like). Things get trickier, however, when your NCS is close to, (bot not exactly) crystallographic (NECS?). Or if you are simply not sure. The best way I can think of to deal with this situation is to road test your Rfree: 1) do something that you know is wrong, like delete a helix, or put some side chains in the wrong place 2) refine with NCS turned on 3) check that Rfree actually goes up 4) un-do the wrong things 5) refine again 6) check that Rfree actually goes down 7) try again with NCS turned off Remembering these timeless words of wisdom: Control, Control, you must learn CONTROL! -Yoda (Jedi Master) -James Holton MAD Scientist On 10/21/2010 8:46 AM, Christina Bourne wrote: Dear all, How would one properly select reflections for R-free in these situations? Presumably if the selection is done in P1 then it mimics twinning or high NCS, such that reflections in both the work and free set will be (potentially?) related by symmetry. -Christina From: Mohinder Pal m...@soton.ac.uk To: CCP4BB@JISCMAIL.AC.UK Sent: Thu, October 21, 2010 7:05:42 AM Subject: [ccp4bb] Regarding space group P1, P21 Dear CCP4BB members, I have solved a protein-drug complex structure in P21212 space group. In this structure, the drug molecule is falling on the two-fold symmetry axis having averaged electron density with 0.5 occupancy. We tried a lot to crystallize this protein-drug complex in different space group but no success so far. I have tried to solve the same data in space group P1 (statistics are fine as I have collected data for 360 degree). The map looks even better with one conformation for a drug. Interestingly, then I reprocessed the same data using imosflm in P21 space group which have penalty 1 compared to 4 for P21212. The structure in P21 is also refining well (with one conformation of the drug compound without symmetry axis at the ligand position). The question is , is it a good practice to solve this structure in P1 and P21 even if the data has higher symmetry? Secondly, I have been advised that I have to be careful to refine structure in P1 as there will be problem regarding observation/parameter ratio if I add too many water molecules. What will be the case if the electron density present for water molecules? I can put restrains to protein structure but I am just curious to know one restrain equals how many observations. I look forward to hear your suggestions. Kind regards, Mohinder Pal *** Jacob Pearson Keller Northwestern University Medical Scientist Training Program Dallos Laboratory F. Searle 1-240 2240 Campus Drive Evanston IL 60208 lab: 847.491.2438 cel: 773.608.9185 email: j-kell...@northwestern.edu ***
Re: [ccp4bb] Regarding space group P1, P21
How you choose to make use of (or ignore) crystallographic symmetry comes down to your view of what constitutes the best model for the sample you're studying. How similar do you believe the molecules are in your crystal? If you describe the model in a higher symmetry space group, you believe that given the information content of the diffraction pattern, the molecules are identical. If you describe it using fewer symmetry operations, you believe the molecules differ in some way. So, how you describe the symmetry of your crystal comes down to determining the simplest model consistent with your experimental observations. Ron On Thu, 21 Oct 2010, Jacob Keller wrote: I have heard many times that it is a black eye to refine in a lower-symmetry spacegroup, but I could never really understand why. The higher symmetry could be considered merely a helpful theoretical lens to improve signal-to-noise, and therefore imposing higher symmetry on the data could be seen as a sort of *leniency* of scientific (or at least empiric) rigor. I think similarly about using discrete spot intensities rather than the whole image--we assume Bragg conditions and neglect certain things about the image between the spots, which is usually valid, but not always. I wonder why it is considered maladroit to refine in a lower spacegroup, then--don't higher spacegroup impose more assumptions than p1? Jacob Keller - Original Message - From: James Holton To: CCP4BB@JISCMAIL.AC.UK Sent: Thursday, October 21, 2010 10:55 AM Subject: Re: [ccp4bb] Regarding space group P1, P21 You pick the Rfree flags in the high-symmetry space group, and then use CAD with OUTLIM SPACE P1 to symmetry-expand them to P1 (or whatever you like). Things get trickier, however, when your NCS is close to, (bot not exactly) crystallographic (NECS?). Or if you are simply not sure. The best way I can think of to deal with this situation is to road test your Rfree: 1) do something that you know is wrong, like delete a helix, or put some side chains in the wrong place 2) refine with NCS turned on 3) check that Rfree actually goes up 4) un-do the wrong things 5) refine again 6) check that Rfree actually goes down 7) try again with NCS turned off Remembering these timeless words of wisdom: Control, Control, you must learn CONTROL! -Yoda (Jedi Master) -James Holton MAD Scientist On 10/21/2010 8:46 AM, Christina Bourne wrote: Dear all, How would one properly select reflections for R-free in these situations? Presumably if the selection is done in P1 then it mimics twinning or high NCS, such that reflections in both the work and free set will be (potentially?) related by symmetry. -Christina __ From: Mohinder Pal m...@soton.ac.uk To: CCP4BB@JISCMAIL.AC.UK Sent: Thu, October 21, 2010 7:05:42 AM Subject: [ccp4bb] Regarding space group P1, P21 Dear CCP4BB members, I have solved a protein-drug complex structure in P21212 space group. In this structure, the drug molecule is falling on the two-fold symmetry axis having averaged electron density with 0.5 occupancy. We tried a lot to crystallize this protein-drug complex in different space group but no success so far. I have tried to solve the same data in space group P1 (statistics are fine as I have collected data for 360 degree). The map looks even better with one conformation for a drug. Interestingly, then I reprocessed the same data using imosflm in P21 space group which have penalty 1 compared to 4 for P21212. The structure in P21 is also refining well (with one conformation of the drug compound without symmetry axis at the ligand position). The question is , is it a good practice to solve this structure in P1 and P21 even if the data has higher symmetry? Secondly, I have been advised that I have to be careful to refine structure in P1 as there will be problem regarding observation/parameter ratio if I add too many water molecules. What will be the case if the electron density present for water molecules? I can put restrains to protein structure but I am just curious to know one restrain equals how many observations. I look forward to hear your suggestions. Kind regards, Mohinder Pal *** Jacob Pearson Keller Northwestern University Medical Scientist Training Program Dallos Laboratory F. Searle 1-240 2240 Campus Drive Evanston IL 60208 lab: 847.491.2438 cel: 773.608.9185 email: j-kell...@northwestern.edu ***
Re: [ccp4bb] Regarding space group P1, P21
Because refining in the (right) higher symmetry space group leads to a better model. On Thu, 2010-10-21 at 11:34 -0500, Jacob Keller wrote: I have heard many times that it is a black eye to refine in a lower-symmetry spacegroup, but I could never really understand why. The higher symmetry could be considered merely a helpful theoretical lens to improve signal-to-noise, and therefore imposing higher symmetry on the data could be seen as a sort of *leniency* of scientific (or at least empiric) rigor. I think similarly about using discrete spot intensities rather than the whole image--we assume Bragg conditions and neglect certain things about the image between the spots, which is usually valid, but not always. I wonder why it is considered maladroit to refine in a lower spacegroup, then--don't higher spacegroup impose more assumptions than p1? Jacob Keller - Original Message - From: James Holton To: CCP4BB@JISCMAIL.AC.UK Sent: Thursday, October 21, 2010 10:55 AM Subject: Re: [ccp4bb] Regarding space group P1, P21 You pick the Rfree flags in the high-symmetry space group, and then use CAD with OUTLIM SPACE P1 to symmetry-expand them to P1 (or whatever you like). Things get trickier, however, when your NCS is close to, (bot not exactly) crystallographic (NECS?). Or if you are simply not sure. The best way I can think of to deal with this situation is to road test your Rfree: 1) do something that you know is wrong, like delete a helix, or put some side chains in the wrong place 2) refine with NCS turned on 3) check that Rfree actually goes up 4) un-do the wrong things 5) refine again 6) check that Rfree actually goes down 7) try again with NCS turned off Remembering these timeless words of wisdom: Control, Control, you must learn CONTROL! -Yoda (Jedi Master) -James Holton MAD Scientist On 10/21/2010 8:46 AM, Christina Bourne wrote: Dear all, How would one properly select reflections for R-free in these situations? Presumably if the selection is done in P1 then it mimics twinning or high NCS, such that reflections in both the work and free set will be (potentially?) related by symmetry. -Christina From: Mohinder Pal m...@soton.ac.uk To: CCP4BB@JISCMAIL.AC.UK Sent: Thu, October 21, 2010 7:05:42 AM Subject: [ccp4bb] Regarding space group P1, P21 Dear CCP4BB members, I have solved a protein-drug complex structure in P21212 space group. In this structure, the drug molecule is falling on the two-fold symmetry axis having averaged electron density with 0.5 occupancy. We tried a lot to crystallize this protein-drug complex in different space group but no success so far. I have tried to solve the same data in space group P1 (statistics are fine as I have collected data for 360 degree). The map looks even better with one conformation for a drug. Interestingly, then I reprocessed the same data using imosflm in P21 space group which have penalty 1 compared to 4 for P21212. The structure in P21 is also refining well (with one conformation of the drug compound without symmetry axis at the ligand position). The question is , is it a good practice to solve this structure in P1 and P21 even if the data has higher symmetry? Secondly, I have been advised that I have to be careful to refine structure in P1 as there will be problem regarding observation/parameter ratio if I add too many water molecules. What will be the case if the electron density present for water molecules? I can put restrains to protein structure but I am just curious to know one restrain equals how many observations. I look forward to hear your suggestions. Kind regards, Mohinder Pal *** Jacob Pearson Keller Northwestern University Medical Scientist Training Program Dallos Laboratory F. Searle 1-240 2240 Campus Drive Evanston IL 60208 lab: 847.491.2438 cel: 773.608.9185 email: j-kell...@northwestern.edu *** -- Edwin Pozharski, PhD, Assistant Professor University of Maryland, Baltimore -- When the
[ccp4bb] Fwd: [ccp4bb] tlsanl *NOW ALSO REFMAC* and phenix TLS results
On Thu, Oct 21, 2010 at 11:42 AM, Ian Tickle ianj...@gmail.com wrote: There are a number of things that don't look right here, both with the Refmac and the Phenix runs: [...] interesting, thanks for these comments. S should not have been symmetrized to actually display the principal axes in molscript, i mean. -bryan
Re: [ccp4bb] **Possible spam**cryoprotetant for 35% Dioxane
Not to forget J. Appl. Cryst. (1996). 29, 584-587[ doi:10.1107/S0021889896004190 ] Glycerol concentrations required for cryoprotection of 50 typical protein crystallization solutions E. F. Garman and E. P. Mitchell Which prompted the McFerrin and Snell work. Also worth checking out (and I apologize if I missed others) Acta Cryst. (2008). D64, 287-301[ doi:10.1107/S0907444907067613 ] Glycerol concentrations required for the successful vitrification of cocktail conditions in a high-throughput crystallization screen R. Kempkes, E. Stofko, K. Lam and E. H. Snell J. Appl. Cryst. (2006). 39, 244-251[ doi:10.1107/S0021889806004717 ] Effects of cryoprotectant concentration and cooling rate on vitrification of aqueous solutions V. Berejnov, N. S. Husseini, O. A. Alsaied and R. E. Thorne Cheers, Eddie Edward Snell Ph.D. Assistant Prof. Department of Structural Biology, SUNY Buffalo, Senior Scientist, Hauptman-Woodward Medical Research Institute 700 Ellicott Street, Buffalo, NY 14203-1102 Phone: (716) 898 8631 Fax: (716) 898 8660 Skype: eddie.snell Email: esn...@hwi.buffalo.edu Telepathy: 42.2 GHz Heisenberg was probably here! -Original Message- From: CCP4 bulletin board [mailto:ccp...@jiscmail.ac.uk] On Behalf Of Eric Larson Sent: Thursday, October 21, 2010 12:14 PM To: CCP4BB@JISCMAIL.AC.UK Subject: Re: [ccp4bb] **Possible spam**cryoprotetant for 35% Dioxane Hi Jerry, A great reference for initial cryocondition searches for many standard crystallization solutions is the tables in: J. Appl. Cryst. (2002). 35, 538-545 [ doi:10.1107/S0021889802009238 ] The development and application of a method to quantify the quality of cryoprotectant solutions using standard area-detector X-ray images M. B. McFerrin and E. H. Snell On the bottom of page 542 it says for 35% dioxane as the precipitant in Hampton Crystal Screen II condition # 4: 25% glycerol, 25% PEG 400, 20% ethylene glycol, 15% propylene glycol (1,2-propanediol) good luck, Eric __ Eric Larson, PhD Biomolecular Structure Center Department of Biochemistry Box 357742 University of Washington Seattle, WA 98195 On Wed, 20 Oct 2010, Jerry McCully wrote: | Dear All; | | We just got some crystals from 35% (v/v) Dioxane. We are going to collect some data soon. | | Does anyone have the experience with the cryoprotectant in this condition? | | Thanks a lot, | | Jerry McCully | | | |
Re: [ccp4bb] **Possible spam**cryoprotetant for 35% Dioxane
And if you are not so into reading papers, you can use this database http://idb.exst.jaxa.jp/db_data/protein/search-e.php? Jürgen P.S. Who wants to write an App for that, wouldn't this be very handy at the beamline ? I take 5% of the income for the idea :-) - Jürgen Bosch Johns Hopkins Bloomberg School of Public Health Department of Biochemistry Molecular Biology Johns Hopkins Malaria Research Institute 615 North Wolfe Street, W8708 Baltimore, MD 21205 Phone: +1-410-614-4742 Lab: +1-410-614-4894 Fax: +1-410-955-3655 http://web.mac.com/bosch_lab/ On Oct 21, 2010, at 1:07 PM, Edward Snell wrote: Not to forget J. Appl. Cryst. (1996). 29, 584-587[ doi:10.1107/S0021889896004190 ] Glycerol concentrations required for cryoprotection of 50 typical protein crystallization solutions E. F. Garman and E. P. Mitchell Which prompted the McFerrin and Snell work. Also worth checking out (and I apologize if I missed others) Acta Cryst. (2008). D64, 287-301[ doi:10.1107/S0907444907067613 ] Glycerol concentrations required for the successful vitrification of cocktail conditions in a high-throughput crystallization screen R. Kempkes, E. Stofko, K. Lam and E. H. Snell J. Appl. Cryst. (2006). 39, 244-251[ doi:10.1107/S0021889806004717 ] Effects of cryoprotectant concentration and cooling rate on vitrification of aqueous solutions V. Berejnov, N. S. Husseini, O. A. Alsaied and R. E. Thorne Cheers, Eddie Edward Snell Ph.D. Assistant Prof. Department of Structural Biology, SUNY Buffalo, Senior Scientist, Hauptman-Woodward Medical Research Institute 700 Ellicott Street, Buffalo, NY 14203-1102 Phone: (716) 898 8631 Fax: (716) 898 8660 Skype: eddie.snell Email: esn...@hwi.buffalo.edu Telepathy: 42.2 GHz Heisenberg was probably here! -Original Message- From: CCP4 bulletin board [mailto:ccp...@jiscmail.ac.uk] On Behalf Of Eric Larson Sent: Thursday, October 21, 2010 12:14 PM To: CCP4BB@JISCMAIL.AC.UK Subject: Re: [ccp4bb] **Possible spam**cryoprotetant for 35% Dioxane Hi Jerry, A great reference for initial cryocondition searches for many standard crystallization solutions is the tables in: J. Appl. Cryst. (2002). 35, 538-545 [ doi:10.1107/S0021889802009238 ] The development and application of a method to quantify the quality of cryoprotectant solutions using standard area-detector X-ray images M. B. McFerrin and E. H. Snell On the bottom of page 542 it says for 35% dioxane as the precipitant in Hampton Crystal Screen II condition # 4: 25% glycerol, 25% PEG 400, 20% ethylene glycol, 15% propylene glycol (1,2-propanediol) good luck, Eric __ Eric Larson, PhD Biomolecular Structure Center Department of Biochemistry Box 357742 University of Washington Seattle, WA 98195 On Wed, 20 Oct 2010, Jerry McCully wrote: | Dear All; | | We just got some crystals from 35% (v/v) Dioxane. We are going to collect some data soon. | | Does anyone have the experience with the cryoprotectant in this condition? | | Thanks a lot, | | Jerry McCully | | | |
Re: [ccp4bb] tlsanl *NOW ALSO REFMAC* and phenix TLS results
It should be straightforward to work out what you need to do to the Phenix output to make it acceptable to TLSANL. All I need is the piece of Phenix documentation that defines the TLS tensors that you are using. Cheers -- Ian On Thu, Oct 21, 2010 at 3:55 PM, Bryan Lepore bryanlep...@gmail.com wrote: documentation then i conclude the TLS protocol in refmac is markedly different from phenix (i know this is not strictly a ccp4 question). cf. : refmac: TLS RANGE 'A 245.' 'A 252.' ALL ORIGIN 14.019 -10.476 -35.068 T 0.4974 0.0372 0.3453 0.0674 0.2984 0.0431 L 21.5463 29.5753 20.7545 39.3304 9.2958 -11.5975 S 0.6432 1.0787 -0.1449 1.2672 2.1065 1.5489 -1.1854 -0.0349 center of reaction : 18.652 -5.646 -40.945 from orth. axes libration perpendicular to TLS plane : 13.256 deg^2 phenix: TLS RANGE 'A 245.' 'A 252.' ALL ORIGIN 14.019 -10.476 -35.068 T 0.2961 0.2533 0.2969 -0.0097 0.0111 0.0205 L 0.0006 0.0005 0.0006 0.0006 -0.0002 -0.0001 S 0.0058 -0.0026 -0.0027 0.0035 -0.0017 0.0011 -0.0052 0.0027 center of reaction : -606.636-734.593 -54.324 from orth. axes libration perpendicular to TLS plane : 0.001 deg^2 (trying to keep this a refmac/tlsanl question) and the results are equivalent - because of factors perhaps in symmetrization of S, because there are no unique choices of these things. if so, only the refmac axes faciliate an intuitive comparison. regards, -bryan p.s: link to original thread phenix board : http://www.phenix-online.org/pipermail/phenixbb/2010-October/004724.html
Re: [ccp4bb] Regarding space group P1, P21
The black eye comes not from the treatment of the observations, but from the treatment of the model. If you want to refine the same model against lower symmetry and/or unmerged data - go right ahead. I think the result will not usually be an improvement, but in some cases this may work around systematic artefacts in the data. What you should _not_ do is replicate the model to produce multiple copies which are then refined as if they were independent. That amounts to doubling/tripling/whatever the number of model parameters. Ethan I think that when you say as if they were independent, you are begging the question. You could say that refining in higher symmetry treats the molecules as if they were the same. Further, it really assumes more to posit that they are the same. Really the crux I think is weighing what benefits one gets from treating the data in different ways. If one can know somehow that the molecules when treated as p1 differ from each other only as a function of experimental noise, there would be no reason to treat them as p1. On the other hand, if somehow a few sidechains became systematically different between molecules in the p1 cell, it *would* make sense to refine in p1, no? (One could imagine an electric field around the crystal upon freezing or whatever.) Jacob Keller
Re: [ccp4bb] Regarding space group P1, P21
Hi Ed, On Thu, Oct 21, 2010 at 12:18:31PM -0400, Ed Pozharski wrote: Let's say I have a ligand on symmetry axes and so it appears in two conformations. If I reduce symmetry, there are two possible scenarios. a. In lower symmetry, ligand still appears in two conformations. Shall use higher symmetry. b. In lower symmetry, ligand appears to be in single conformation (this is what Mohinder says, if I am not mistaken). In this case, the true symmetry is lower, and it is simply overwhelmed by the fact that most of the structure (but not all) obeys higher symmetry. I think I understand what you're getting at: you have a lower symmetry with a NCS axis that is basically perfectly aligned with the corresponding crystallographic axis in the higher symmetry spacegroup. And the only part of the model not obeying this NCS is the ligand. But then what about a water on a special position (2-fold with occ=0.5)? If I remove that 2-fold from my spacegroup symmetry and refine I get ... a single water with occupancy 1.0 ... or 2 waters with occupancy 0.5? Hmmm, diffcult to decide on the true spacegroup here ;-) So it all depends * how clear the difference between high-symmetry/double conformation and low-symmetry/single-conformation is * how symmetrical the ligand is * how the refinement in the lower-symmetry spacegroup is done - since there is a real danger (in case it is the high-symmetry spacegroup after all) that because of model bias and poorer (independent observations)/parameter ratio what seems like a clear single conformation is difficult to confirm. I recall Bruce Foxman describing a b) case (I am sure there is more than one example) for a small molecule crystal, where a single heavy atom had higher symmetry than the rest of the molecule. There is a recent nice example of a very interesting symmetry/disorder siuation by Yves Muller: 2xgc. It took some time for me to get my head around what it is in the PDB file and what it means ... but it's very neat! Cheers Clemens -- *** * Clemens Vonrhein, Ph.D. vonrhein AT GlobalPhasing DOT com * * Global Phasing Ltd. * Sheraton House, Castle Park * Cambridge CB3 0AX, UK *-- * BUSTER Development Group (http://www.globalphasing.com) ***
Re: [ccp4bb] Regarding space group P1, P21
On Thu, 2010-10-21 at 12:58 -0500, Jacob Keller wrote: On the other hand, if somehow a few sidechains became systematically different between molecules in the p1 cell, it *would* make sense to refine in p1 And sometimes (but rarely) such differences become detectable at high resolution (e.g. Kraut et al., PLOS Biology, 4:501). -- I'd jump in myself, if I weren't so good at whistling. Julian, King of Lemurs
Re: [ccp4bb] Regarding space group P1, P21
Hi, I think that when you say as if they were independent, you are begging the question. You could say that refining in higher symmetry treats the molecules as if they were the same. Further, it really assumes more to posit that they are the same. But we're still talking about crystals, right? The whole reason for trying to crystallise our proteins/DNA/RNA is because we ideally want a perfect arrangement of molecules. So taking as a starting hypotheses the conservative approach that if the data really looks like P21 it probably is P21 seems a good idea to me. To me it is more a case of refining in lower symmetry treats the molecules as if they were different when initially we don't have an indication for it (from data processing). Unless we take the fact that P1 will always give us lower merging R-factors and better indexing scores as indication that actually all our crystals are always P1 ... which they well might be, but probably not within our experimental error. Really the crux I think is weighing what benefits one gets from treating the data in different ways. If one can know somehow that the molecules when treated as p1 differ from each other only as a function of experimental noise, there would be no reason to treat them as p1. True: but how do you judge that those differences are within or outside of experimental noise? On the other hand, if somehow a few sidechains became systematically different between molecules in the p1 cell, it *would* make sense to refine in p1, no? What if by refining in P1 the parametrisation makes those side-chains different in the first place? A poorly defined Lys side-chain suddenly becomes two significantly different poorly defined side-chain? Cheers Clemens -- *** * Clemens Vonrhein, Ph.D. vonrhein AT GlobalPhasing DOT com * * Global Phasing Ltd. * Sheraton House, Castle Park * Cambridge CB3 0AX, UK *-- * BUSTER Development Group (http://www.globalphasing.com) ***
Re: [ccp4bb] Regarding space group P1, P21
On Thu, 2010-10-21 at 18:59 +0100, Clemens Vonrhein wrote: I think I understand what you're getting at: you have a lower symmetry with a NCS axis that is basically perfectly aligned with the corresponding crystallographic axis in the higher symmetry spacegroup. And the only part of the model not obeying this NCS is the ligand. precisely But then what about a water on a special position (2-fold with occ=0.5)? If I remove that 2-fold from my spacegroup symmetry and refine I get ... a single water with occupancy 1.0 ... or 2 waters with occupancy 0.5? Hmmm, diffcult to decide on the true spacegroup here ;-) water is symmetrical (no hydrogens, please), shall use the higher symmetry So it all depends * how clear the difference between high-symmetry/double conformation and low-symmetry/single-conformation is Hard to put a specific number on it. I'd inspect the density and play Potter Stewart. * how symmetrical the ligand is same deal as with water * how the refinement in the lower-symmetry spacegroup is done - since there is a real danger (in case it is the high-symmetry spacegroup after all) that because of model bias and poorer (independent observations)/parameter ratio what seems like a clear single conformation is difficult to confirm. Absolutely true. As we discussed before, restraining protein copies is a must as well as maybe some bias removal. Cheers, Ed. -- I'd jump in myself, if I weren't so good at whistling. Julian, King of Lemurs
Re: [ccp4bb] Regarding space group P1, P21
But we're still talking about crystals, right? The whole reason for trying to crystallise our proteins/DNA/RNA is because we ideally want a perfect arrangement of molecules. So taking as a starting hypotheses the conservative approach that if the data really looks like P21 it probably is P21 seems a good idea to me. if the data really looks like P21-- what are the criteria for that? For example, I believe p1 can have good-as-perfect 90deg angles, no? And also equal cell dimensions? So I don't think you will be able to tell from the positions of the spots on the detector, necessarily. Also, would it not be more rigorous to say I can gain a lot by assuming these molecules are in p21? Look, nobody thinks that every molecule in the crystal is identical, so that is truly a convenient assumption. The symmetry, I think, is a similar assumption at a different level. By the way, I have always wondered whether anybody has looked into the degree of intermolecular differences possible given all of the parameters in our crystallographic models. In other words, would a microscopic observer look at the molecules in the crystal and see what looks like a crowd from a NYC street, or something more like an army formation? How much variety is there at the molecular lever, I wonder? True: but how do you judge that those differences are within or outside of experimental noise? Agreed! What if by refining in P1 the parametrisation makes those side-chains different in the first place? A poorly defined Lys side-chain suddenly becomes two significantly different poorly defined side-chain? I don't know--depends on last question I think. Jacob *** Jacob Pearson Keller Northwestern University Medical Scientist Training Program Dallos Laboratory F. Searle 1-240 2240 Campus Drive Evanston IL 60208 lab: 847.491.2438 cel: 773.608.9185 email: j-kell...@northwestern.edu ***
Re: [ccp4bb] FOM: Phaser vs SigmaA
Thank you very much. I am in total agreement with you that HL coefficients are better description of the phase probability and in fact I have been using the HL coeffs all these times for phase combination and density modification. My earlier post was just to make it easy for a comparison between weights calculated by sigmaA and Phaser. If I take HL coeffs from Phaser and calculate FOM, I will also get the same unusually low FOM. (I am not familiar with how to compare HL coeffs directly but based on the HL equation, the higher the coeff values for the same phase set, the higher the phase probability. The HL coeffs from Phaser are much smaller (A=0.3, B=0.28, C,D=0) than HL coeffs converted from SigmaA (A=0.95, B=0.88). The values are mean abs. from the mtz files.) The reason I am curious about the low FOM (or HL coeffs as you prefer) from Phaser is because I am trying to compare phase combination results between the partial structure and low resolution experimental phases. I use poly alanine alpha-helices as a partial structure and apparently when using HL coeffs calculated from sigmaA or Refmac results, the model phases contribute quite a large model bias after phase combination which also get carried through density modification. The model bias seems less (obviously because of lower weight from the model) when I use result from Phaser and and the map after density modification somehow seems more satisfying. BTW, I am testing a method to solve previously determined strutures starting with experimental low resolution phases and that is why I was curious to know how Phaser gives so much different weight and HL coeffs from sigmaA. I know that you can change weight by using a scale factor during phase combination (also available in CCP4 clipper utility) or bluring the phase probability. The easiest would be to arbitarily scale it down by half. But if anybody knows a good criteria to properly down weight phase probability, please let me know. Many thanks, George Wisedchaisri On Thu, 21 Oct 2010, Clemens Vonrhein wrote: Hi George, On Wed, Oct 20, 2010 at 04:58:34PM -0700, Goragot Wisedchaisri wrote: Hi, I have helices that I did rigid body refinement with Phaser (after phased rotation and phased translation in Molrep). I compare FOM output by Phaser to the FOM computed by sigmaA using the Phaser refined coordinates and found that FOM from Phaser is only about half (~0.25) of FOM from SigmaA (~0.5). What do you need FOM values for - apart from just looking at them? You don't need them for calculating maps since both SIGMAA and Phaser (I assume) output map-coefficients directly (an amplitude and weight). And you don't need them in refinement since both programs probably output Hendrickson-Lattmann coefficients - which are a much better description of phase probability. Density-modification: same thing. I could just use SigmaA or do refinement in Refmac but I have to say that I like the low FOM from Phaser because model bias seem to be much less after density modification. Are you using the FOM columns in density modification? Why? Most modern programs will allow you input of Hendrickson-Lattmann coefficients (and also output those). And the initial map can be calculated with the map coefficients anyway. It also saves me from having to blur the phase probability distribution in order to down weight FOM when FOM is too high. FOM columns in a MTZ file are maybe useful to calculate statistics versus resolution ... but nearly everything you would do with FOM (attach it to an amplitude and phase) can be done much better with Hendrickson-Lattmann coefficients. E.g. running DM after a MR solution, I would use LABIN FP=F SIGFP=SIGF PHIO=PHWT FOMO=FOM - HLA=HLA HLB=HLB HLC=HLC HLD=HLD - FDM=FWT PHIDM=PHWT It still reads the FOM column - but only to analyse it against resolution to determine a good phase-extension scheme. Internally, it will always use the HLA-D values ... afaik. Cheers Clemens -- *** * Clemens Vonrhein, Ph.D. vonrhein AT GlobalPhasing DOT com * * Global Phasing Ltd. * Sheraton House, Castle Park * Cambridge CB3 0AX, UK *-- * BUSTER Development Group (http://www.globalphasing.com) ***
Re: [ccp4bb] Regarding space group P1, P21
Hi Clemens, Sorry to be picky and start the 'definition game' over again, but 'Miller indices' are strictly not the numbers that index X-ray reflections that everyone is familiar with (whether observed or not!). Miller indices were introduced in 1839 by the British mineralogist William Hallowes Miller (it says in WIkipedia) as a way of describing the direction of the perpendicular to the plane faces that he observed on mineral crystals. A condition is that no common denominator is possible, since it defines only the direction of a vector; its magnitude has no relevance in this context. So you can have Miller indices (1,0,0), (1,2,0), (1,2,3) etc but you can't have (2,0,0), (3,0,0), {2,4,0), (3,6,9) etc., or at least (1,0,0) means exactly the same thing as (2,0,0) etc. You can multiply the MiIler index vector by -1: this indicates the opposite face of the crystal. Imagine what an electron density map would look like if you only collected intensities at the Miller indices! Miller's observation of the plane faces of mineral crystals occurred 73 years before the discovery in 1912 of X-ray diffraction by Max Laue in Munich (he became Max von Laue in 1913 when his father was raised to the nobility), for which Laue received the Nobel Prize in Physics in 1914. Laue explained diffraction by means of the 'Laue equations' which contain 3 integers corresponding exactly to the indices we are all familiar with. I prefer to call them 'reflection indices', though strictly I suppose we should be calling them 'Laue indices'. Almost immediately after Laue's discovery, William Lawrence Bragg in Cambridge devised what we now know as Bragg's Law, wherein the factor 'n' relates the Miller indices to the Laue indices; thus the reflection with indices (nh,nk,nl) is the n'th order of diffraction from the set of crystal planes with Miller indices (h,k,l). Bragg also received the physics Nobel prize jointly with his father William Henry Bragg in the following year, 1915, for their determination of the crystal structures of NaCl, ZnS and diamond. Cheers -- Ian On Thu, Oct 21, 2010 at 4:57 PM, Clemens Vonrhein vonrh...@globalphasing.com wrote: Hi Herman, On Thu, Oct 21, 2010 at 05:31:51PM +0200, herman.schreu...@sanofi-aventis.com wrote: If you process your data in a lower symmetry space group, you will have more unique reflections, since reflections which are related by the higher symmetry will be avaraged during scaling in a higher symmetry space group (i.e. a 2fold or 3fold axis), while in lower symmetry space groups they will not. So the observation to parameter ratio stays the same and is only depending on resolution and solvent content. True - if you count Miller indices as observations. But if you think about information content than probably not (as you discuss below). The question one has to ask of course is: are these reflections really different, or are they the same only not averaged? Yes - by merging we're getting better data (better error estimate on the intensity due to higher multiplicity). So there isn't really independent information in 50% of the reflections if e.g. going from P21 to P1 - we've only increased the noise because the multiplicity of each reflection has been reduced. In the latter case, you have more reflections, but not more information. As Ed mentions, using tight NCS restraints would in this case mimick the crystallographic symmetry. Apart from the (good) NCS argument, one could go even further: We could also just collect 36000 degree of data on a 7A Lysozyme crystal and refine against completely unmerged data. After all, why should we stop at removing only the some symmetry operators from our data merging ... lets get rid of all of them including th x,y,z operator and use unmerged data. Then we could refine Lysozyme with anisotropic hydrogens and no restraints against 7A data since we have a huge number of 'observations' ... right? But seriously: there is a difference in having reflections (H, K, L) and independent data (I, SIGI). Maybe we should talk more about (independent observations)/parameters ratio in the same way we look at depdencies of parameters (e.g. restraints on Bfactors etc). Cheers Clemens -- *** * Clemens Vonrhein, Ph.D. vonrhein AT GlobalPhasing DOT com * * Global Phasing Ltd. * Sheraton House, Castle Park * Cambridge CB3 0AX, UK *-- * BUSTER Development Group (http://www.globalphasing.com) ***
Re: [ccp4bb] Regarding space group P1, P21
Well no, I never did during my crystallography training: it seems to be a change of definition that's occurred fairly recently, without recognition of the fact that the original definition is still in use, particularly in mineralogy of course, where, unlike often is the case with protein crystals, you can usually see the crystal faces with the naked eye! I'm thinking particularly of this site that Bernhard recently pointed out: http://news.nationalgeographic.com/news/bigphotos/82948445.html I remember the time when we did actually measure the faces of a crystal (small molecule, not protein) and determine their Miller indices, in order to calculate the absorption correction (no doubt Shel-X still allows you to do it that way!). So it would have been a little confusing to call Miller indices and reflection/Laue indices by the same name! Cheers -- Ian On Thu, Oct 21, 2010 at 8:28 PM, Jacob Keller j-kell...@fsm.northwestern.edu wrote: I like your more-accurate definition, but practically speaking, doesn't everyone call hkl Miller indices? Jacob - Original Message - From: Ian Tickle ianj...@gmail.com To: CCP4BB@JISCMAIL.AC.UK Sent: Thursday, October 21, 2010 2:00 PM Subject: Re: [ccp4bb] Regarding space group P1, P21 Hi Clemens, Sorry to be picky and start the 'definition game' over again, but 'Miller indices' are strictly not the numbers that index X-ray reflections that everyone is familiar with (whether observed or not!). Miller indices were introduced in 1839 by the British mineralogist William Hallowes Miller (it says in WIkipedia) as a way of describing the direction of the perpendicular to the plane faces that he observed on mineral crystals. A condition is that no common denominator is possible, since it defines only the direction of a vector; its magnitude has no relevance in this context. So you can have Miller indices (1,0,0), (1,2,0), (1,2,3) etc but you can't have (2,0,0), (3,0,0), {2,4,0), (3,6,9) etc., or at least (1,0,0) means exactly the same thing as (2,0,0) etc. You can multiply the MiIler index vector by -1: this indicates the opposite face of the crystal. Imagine what an electron density map would look like if you only collected intensities at the Miller indices! Miller's observation of the plane faces of mineral crystals occurred 73 years before the discovery in 1912 of X-ray diffraction by Max Laue in Munich (he became Max von Laue in 1913 when his father was raised to the nobility), for which Laue received the Nobel Prize in Physics in 1914. Laue explained diffraction by means of the 'Laue equations' which contain 3 integers corresponding exactly to the indices we are all familiar with. I prefer to call them 'reflection indices', though strictly I suppose we should be calling them 'Laue indices'. Almost immediately after Laue's discovery, William Lawrence Bragg in Cambridge devised what we now know as Bragg's Law, wherein the factor 'n' relates the Miller indices to the Laue indices; thus the reflection with indices (nh,nk,nl) is the n'th order of diffraction from the set of crystal planes with Miller indices (h,k,l). Bragg also received the physics Nobel prize jointly with his father William Henry Bragg in the following year, 1915, for their determination of the crystal structures of NaCl, ZnS and diamond. Cheers -- Ian On Thu, Oct 21, 2010 at 4:57 PM, Clemens Vonrhein vonrh...@globalphasing.com wrote: Hi Herman, On Thu, Oct 21, 2010 at 05:31:51PM +0200, herman.schreu...@sanofi-aventis.com wrote: If you process your data in a lower symmetry space group, you will have more unique reflections, since reflections which are related by the higher symmetry will be avaraged during scaling in a higher symmetry space group (i.e. a 2fold or 3fold axis), while in lower symmetry space groups they will not. So the observation to parameter ratio stays the same and is only depending on resolution and solvent content. True - if you count Miller indices as observations. But if you think about information content than probably not (as you discuss below). The question one has to ask of course is: are these reflections really different, or are they the same only not averaged? Yes - by merging we're getting better data (better error estimate on the intensity due to higher multiplicity). So there isn't really independent information in 50% of the reflections if e.g. going from P21 to P1 - we've only increased the noise because the multiplicity of each reflection has been reduced. In the latter case, you have more reflections, but not more information. As Ed mentions, using tight NCS restraints would in this case mimick the crystallographic symmetry. Apart from the (good) NCS argument, one could go even further: We could also just collect 36000 degree of data on a 7A Lysozyme crystal and refine against completely unmerged data. After all, why should we stop at removing only the some symmetry operators
Re: [ccp4bb] Regarding space group P1, P21
On Thursday, October 21, 2010 11:38:55 am Jacob Keller wrote: if the data really looks like P21-- what are the criteria for that? This is a straightforward statistical question. In testing for a possible 2-fold, you want to know: Do two random reflections related by the putative 2-fold agree with each other better than two random reflections not related by the putative 2-fold? To make this test less sensitive to scaling, one can formulate it as a correlation coefficient. Have a look at the paper describing `pointless`. P Evans (2006), Acta Cryst. D62: 72-82 Testing the for systematic absences indicating a screw axis can also be phrased as a statistical test, although generally there are a relatively small number of putative absences to inspect so the test is not all that strong. Ethan I believe p1 can have good-as-perfect 90deg angles, no? Correct. The cell angles don't really enter into it. And also equal cell dimensions? So I don't think you will be able to tell from the positions of the spots on the detector, necessarily. Also, would it not be more rigorous to say I can gain a lot by assuming these molecules are in p21? Look, nobody thinks that every molecule in the crystal is identical, so that is truly a convenient assumption. The symmetry, I think, is a similar assumption at a different level. By the way, I have always wondered whether anybody has looked into the degree of intermolecular differences possible given all of the parameters in our crystallographic models. In other words, would a microscopic observer look at the molecules in the crystal and see what looks like a crowd from a NYC street, or something more like an army formation? How much variety is there at the molecular lever, I wonder? True: but how do you judge that those differences are within or outside of experimental noise? Agreed! What if by refining in P1 the parametrisation makes those side-chains different in the first place? A poorly defined Lys side-chain suddenly becomes two significantly different poorly defined side-chain? I don't know--depends on last question I think. Jacob *** Jacob Pearson Keller Northwestern University Medical Scientist Training Program Dallos Laboratory F. Searle 1-240 2240 Campus Drive Evanston IL 60208 lab: 847.491.2438 cel: 773.608.9185 email: j-kell...@northwestern.edu *** -- Ethan A Merritt Biomolecular Structure Center, K-428 Health Sciences Bldg University of Washington, Seattle 98195-7742
Re: [ccp4bb] Regarding space group P1, P21
It's more complicated than that, since the tricky thing is to distinguish between reflections related eg by a putative crystallographic two-fold and by a parallel non-crystallographic two-fold, which would give very similar intensity relationships. Pointless does try to score these alternative models, but it is not fool-proof. In the end, the best test is probably comparing refinements in different space groups (as is done by the Andrey Lebedev's Zanuda program, on the York University server), though it seems to me that in the limit you can't tell: how close does a non-crystallographic axis have to be to a crystal direction to be crystallographic, 1degree, 0.1 degrees, 0.01degrees? Phil Evans (incidentally, the algorithms used in Pointless are described in a paper due to appear in the Acta Cryst. D volume from the 2010 CCP4 Study Weekend, probably early next year. But I don't really know how best to calculate the probabilities) On 21 Oct 2010, at 21:03, Ethan Merritt wrote: On Thursday, October 21, 2010 11:38:55 am Jacob Keller wrote: if the data really looks like P21-- what are the criteria for that? This is a straightforward statistical question. In testing for a possible 2-fold, you want to know: Do two random reflections related by the putative 2-fold agree with each other better than two random reflections not related by the putative 2-fold? To make this test less sensitive to scaling, one can formulate it as a correlation coefficient. Have a look at the paper describing `pointless`. P Evans (2006), Acta Cryst. D62: 72-82 Testing the for systematic absences indicating a screw axis can also be phrased as a statistical test, although generally there are a relatively small number of putative absences to inspect so the test is not all that strong. Ethan I believe p1 can have good-as-perfect 90deg angles, no? Correct. The cell angles don't really enter into it. And also equal cell dimensions? So I don't think you will be able to tell from the positions of the spots on the detector, necessarily. Also, would it not be more rigorous to say I can gain a lot by assuming these molecules are in p21? Look, nobody thinks that every molecule in the crystal is identical, so that is truly a convenient assumption. The symmetry, I think, is a similar assumption at a different level. By the way, I have always wondered whether anybody has looked into the degree of intermolecular differences possible given all of the parameters in our crystallographic models. In other words, would a microscopic observer look at the molecules in the crystal and see what looks like a crowd from a NYC street, or something more like an army formation? How much variety is there at the molecular lever, I wonder? True: but how do you judge that those differences are within or outside of experimental noise? Agreed! What if by refining in P1 the parametrisation makes those side-chains different in the first place? A poorly defined Lys side-chain suddenly becomes two significantly different poorly defined side-chain? I don't know--depends on last question I think. Jacob *** Jacob Pearson Keller Northwestern University Medical Scientist Training Program Dallos Laboratory F. Searle 1-240 2240 Campus Drive Evanston IL 60208 lab: 847.491.2438 cel: 773.608.9185 email: j-kell...@northwestern.edu *** -- Ethan A Merritt Biomolecular Structure Center, K-428 Health Sciences Bldg University of Washington, Seattle 98195-7742
Re: [ccp4bb] update--degradation of MBP fusion protein
Hi Jerry, Try parameters that slow the rate of protein production, such as expressing at lower temperature, using less IPTG/inducing agent, or using a weaker promoter. Good luck, Ho