[ccp4bb] Postdoctoral and Ph.D. opportunities at the Max Planck Institute for Medical Research, Heidelberg
Postdoctoral Position and PhD positions in Structural Biology at the Max Planck Institute for Medical Research, Heidelberg. The Meinhart lab in the Department of Biomolecular Mechanisms seeks to recruit outstanding postdoctoral scientists and PhD students with experience and/or interest in mechanistic studies of either RNA processing machines (maturation of RNA 3’-ends) or macromolecular complexes that lead to programmed cell death and virulence in pathogenic bacteria (by poisoning bacterial cell wall synthesis). Depending on previous experience and scientific interest, successful candidates will contribute to the analysis of these macromolecular assemblies using structural (X-ray crystallography), biophysical / biochemical methods (fluorescence-based assays, isothermal titration calorimetry, analytical ultracentrifucation, etc.) and molecular biology (in vivo validation of structure-based functional hypotheses). The position provides the opportunity for broad training in protein expression, crystallization using high-throughput pipelines, kinetics, spectroscopy, genetics, and molecular biology. For more information, please check the laboratory website (http://www.mpimf-heidelberg.mpg.de/groups/rna_processing). The Meinhart lab is embedded in the Department of Biomolecular Mechnisms offering a unique, multi-disciplinary and international environment allowing to develop an multifaceted research experience which provides an ideal starting-point for a successful scientific career (for further information see: http://www.mpimf-heidelberg.mpg.de/departments/biomolecular_mechanisms). Heidelberg is one of the top centers for biomedical research in Germany, and graduate students will have access to several different Ph.D. programs. Applicants for postdoctoral positions should possess/ expect to obtain a PhD in natural sciences and have at least one peer-reviewed publication as a first author. They should have experience in aspects of a structural biology pipeline, for example protein production and purification, X-ray crystallography etc. Past experience with complementary biophysical techniques (ITC, AUC, CD etc.) and molecular biology is desirable. Please send CV, a brief summary of research experience and interest, contact address of three referees to bmm.recruitm...@mpimf-heidelberg.mpg.de referring to “AM_Recruitment_2012#1” Applicants for PhD positions should have completed their master in natural sciences and have background in biochemistry, structural biology or molecular biology. Experience with X-ray crystallography is not required but a dedication to biophysics and structural biology techniques is desirable. Please send CV, brief summary of previous research experience and contact address of two referees to bmm.recruitm...@mpimf-heidelberg.mpg.de referring to “AM_Recruitment_2012#2” Applications will be considered until the positions are filled. Best regards, Anton Meinhart _ Dr. Anton Meinhart Department of Biomolecular Mechanisms Max-Planck-Institute for Medical Research Jahnstraße 29 69120 Heidelberg GERMANY phone + 49 6221 / 486505 e-mail: anton.meinh...@mpimf-heidelberg.mpg.de
Re: [ccp4bb] Reasoning for Rmeas or Rpim as Cutoff
Hi, Here are a couple of links on the idea of judging resolution by a type of cross-validation with data not used in refinement: Ling et al, 1998: http://pubs.acs.org/doi/full/10.1021/bi971806n Brunger et al, 2008: http://journals.iucr.org/d/issues/2009/02/00/ba5131/index.html (cites earlier relevant papers from Brunger's group) Best wishes, Randy Read On 30 Jan 2012, at 07:09, arka chakraborty wrote: Hi all, In the context of the above going discussion can anybody post links for a few relevant articles? Thanks in advance, ARKO On Mon, Jan 30, 2012 at 3:05 AM, Randy Read rj...@cam.ac.uk wrote: Just one thing to add to that very detailed response from Ian. We've tended to use a slightly different approach to determining a sensible resolution cutoff, where we judge whether there's useful information in the highest resolution data by whether it agrees with calculated structure factors computed from a model that hasn't been refined against those data. We first did this with the complex of the Shiga-like toxin B-subunit pentamer with the Gb3 trisaccharide (Ling et al, 1998). From memory, the point where the average I/sig(I) drops below 2 was around 3.3A. However, we had a good molecular replacement model to solve this structure and, after just carrying out rigid-body refinement, we computed a SigmaA plot using data to the edge of the detector (somewhere around 2.7A, again from memory). The SigmaA plot dropped off smoothly to 2.8A resolution, with values well above zero (indicating significantly better than random agreement), then dropped suddenly. So we chose 2.8A as the cutoff. Because there were four pentamers in the asymmetric unit, we could then use 20-fold NCS averaging, which gave a fantastic map. In this case, the averaging certainly helped to pull out something very useful from a very weak signal, because the maps weren't nearly as clear at lower resolution. Since then, a number of other people have applied similar tests. Notably, Axel Brunger has done some careful analysis to show that it can indeed be useful to take data beyond the conventional limits. When you don't have a great MR model, you can do something similar by limiting the resolution for the initial refinement and rebuilding, then assessing whether there's useful information at higher resolution by using the improved model (which hasn't seen the higher resolution data) to compute Fcalcs. By the way, it's not necessary to use a SigmaA plot -- the correlation between Fo and Fc probably works just as well. Note that, when the model has been refined against the lower resolution data, you'll expect a drop in correlation at the resolution cutoff you used for refinement, unless you only use the cross-validation data for the resolution range used in refinement. - Randy J. Read Department of Haematology, University of Cambridge Cambridge Institute for Medical ResearchTel: +44 1223 336500 Wellcome Trust/MRC Building Fax: +44 1223 336827 Hills RoadE-mail: rj...@cam.ac.uk Cambridge CB2 0XY, U.K. www-structmed.cimr.cam.ac.uk On 29 Jan 2012, at 17:25, Ian Tickle wrote: Jacob, here's my (personal) take on this: The data quality metrics that everyone uses clearly fall into 2 classes: 'consistency' metrics, i.e. Rmerge/meas/pim and CC(1/2) which measure how well redundant observations agree, and signal/noise ratio metrics, i.e. mean(I/sigma) and completeness, which relate to the information content of the data. IMO the basic problem with all the consistency metrics is that they are not measuring the quantity that is relevant to refinement and electron density maps, namely the information content of the data, at least not in a direct and meaningful way. This is because there are 2 contributors to any consistency metric: the systematic errors (e.g. differences in illuminated volume and absorption) and the random errors (from counting statistics, detector noise etc.). If the data are collected with sufficient redundancy the systematic errors should hopefully largely cancel, and therefore only the random errors will determine the information content. Therefore the systematic error component of the consistency measure (which I suspect is the biggest component, at least for the strong reflections) is not relevant to measuring the information content. If the consistency measure only took into account the random error component (which it can't), then it would be essentially be a measure of information content, if only indirectly (but then why not simply use a direct measure such as the signal/noise ratio?). There are clearly at least 2 distinct problems with Rmerge, first it's including systematic errors in its measure of consistency, second it's not invariant with
[ccp4bb] to show multiple sequence alignment with sec. str.
Dear All, Is there any module in CCP4/ other related software/servers which can show a multiple alignment of homologous sequences from a protein family, together with their secondary structures? Thanks in advance, regards, sreetama
Re: [ccp4bb] Reasoning for Rmeas or Rpim as Cutoff
Hi Randy - thank you for a very interesting reminder to old literature. I'm intrigued: how come this apparently excellent idea has not become standard best practice in the 14 years since it was published? phx On 30/01/2012 09:40, Randy Read wrote: Hi, Here are a couple of links on the idea of judging resolution by a type of cross-validation with data not used in refinement: Ling et al, 1998: http://pubs.acs.org/doi/full/10.1021/bi971806n Brunger et al, 2008: http://journals.iucr.org/d/issues/2009/02/00/ba5131/index.html (cites earlier relevant papers from Brunger's group) Best wishes, Randy Read On 30 Jan 2012, at 07:09, arka chakraborty wrote: Hi all, In the context of the above going discussion can anybody post links for a few relevant articles? Thanks in advance, ARKO On Mon, Jan 30, 2012 at 3:05 AM, Randy Read rj...@cam.ac.uk mailto:rj...@cam.ac.uk wrote: Just one thing to add to that very detailed response from Ian. We've tended to use a slightly different approach to determining a sensible resolution cutoff, where we judge whether there's useful information in the highest resolution data by whether it agrees with calculated structure factors computed from a model that hasn't been refined against those data. We first did this with the complex of the Shiga-like toxin B-subunit pentamer with the Gb3 trisaccharide (Ling et al, 1998). From memory, the point where the average I/sig(I) drops below 2 was around 3.3A. However, we had a good molecular replacement model to solve this structure and, after just carrying out rigid-body refinement, we computed a SigmaA plot using data to the edge of the detector (somewhere around 2.7A, again from memory). The SigmaA plot dropped off smoothly to 2.8A resolution, with values well above zero (indicating significantly better than random agreement), then dropped suddenly. So we chose 2.8A as the cutoff. Because there were four pentamers in the asymmetric unit, we could then use 20-fold NCS averaging, which gave a fantastic map. In this case, the averaging certainly helped to pull out something very useful from a very weak signal, because the maps weren't nearly as clear at lower resolution. Since then, a number of other people have applied similar tests. Notably, Axel Brunger has done some careful analysis to show that it can indeed be useful to take data beyond the conventional limits. When you don't have a great MR model, you can do something similar by limiting the resolution for the initial refinement and rebuilding, then assessing whether there's useful information at higher resolution by using the improved model (which hasn't seen the higher resolution data) to compute Fcalcs. By the way, it's not necessary to use a SigmaA plot -- the correlation between Fo and Fc probably works just as well. Note that, when the model has been refined against the lower resolution data, you'll expect a drop in correlation at the resolution cutoff you used for refinement, unless you only use the cross-validation data for the resolution range used in refinement. - Randy J. Read Department of Haematology, University of Cambridge Cambridge Institute for Medical ResearchTel: +44 1223 336500 tel:%2B44%201223%20336500 Wellcome Trust/MRC Building Fax: +44 1223 336827 tel:%2B44%201223%20336827 Hills Road E-mail: rj...@cam.ac.uk mailto:rj...@cam.ac.uk Cambridge CB2 0XY, U.K. www-structmed.cimr.cam.ac.uk http://www-structmed.cimr.cam.ac.uk/ On 29 Jan 2012, at 17:25, Ian Tickle wrote: Jacob, here's my (personal) take on this: The data quality metrics that everyone uses clearly fall into 2 classes: 'consistency' metrics, i.e. Rmerge/meas/pim and CC(1/2) which measure how well redundant observations agree, and signal/noise ratio metrics, i.e. mean(I/sigma) and completeness, which relate to the information content of the data. IMO the basic problem with all the consistency metrics is that they are not measuring the quantity that is relevant to refinement and electron density maps, namely the information content of the data, at least not in a direct and meaningful way. This is because there are 2 contributors to any consistency metric: the systematic errors (e.g. differences in illuminated volume and absorption) and the random errors (from counting statistics, detector noise etc.). If the data are collected with sufficient redundancy the systematic errors should hopefully largely cancel, and therefore only the random errors will determine the information content. Therefore the systematic error component of the consistency measure (which I suspect is
[ccp4bb] Kinase crystallization
Dear all, I am trying to crystallize a protein kinase without any success. I suspect about its characteristic catalytic loop. I have already prepared different constructs, different expression vectors, and different mutant proteins (pseudo-phosphorylated, active, inactive?). I have also tested some ligands as ANPpnp (non hidrolizable nucleotide), ADP (product) and manganese (cofactor). I am thinking in trying to cocrystallize it with a general kinase substrate (a peptide, a small molecule...). Does any one have any experience or suggestion? Thanks in advance. Sincerely, Antonio
Re: [ccp4bb] to show multiple sequence alignment with sec. str.
Hi Sreetama, you can use STRAP for that: http://3d-alignment.eu/. It allows you to do multible sequence alignments and use various algorithms to predict secondary structure or display secondary structure assignments of PDB entries. Cheers Florian Am 30.01.2012 um 11:02 schrieb sreetama das: Dear All, Is there any module in CCP4/ other related software/servers which can show a multiple alignment of homologous sequences from a protein family, together with their secondary structures? Thanks in advance, regards, sreetama - Dr. Florian Brückner Biomolecular Research Laboratory OFLG/102 Paul Scherrer Institut CH-5232 Villigen PSI Switzerland Tel.: +41-(0)56-310-2332 Email: florian.brueck...@psi.ch
Re: [ccp4bb] Ligand chirality error
Do remember you can assign chirality as both.. This can be useful as otherwise the refinement programs force the requested chirality and espec at low resolution it can be hard to see any indication of error.. .. Eleanor On 01/27/2012 09:06 AM, herman.schreu...@sanofi.com wrote: Dear Debajyoti, The way I check the chirality is to compare the fitted compound with the structural formula. In coot, I rotate the ligand such that it has the same orientation as in the formula and check that the out of plane group goes in the right direction. However, it happens quite often that the chirality of the bound compound is not the chirality the chemist thinks it has, or the chirality is not known. In that case, I refine both enantiomers and look which one fits best. Things to look for are distorted bond angles, poor fitting and small positive and negative blobs of difference density (green and red blobs in coot with default settings). To change the chirality, you have to edit the cif dictionary which contains the description of your compound. In the dictionaries I use you have to change the _chem_comp_chir.volume.sign from positiv to negativ or vice versa. Good luck! Herman From: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] On Behalf Of Debajyoti Dutta Sent: Friday, January 27, 2012 8:25 AM To: CCP4BB@JISCMAIL.AC.UK Subject: [ccp4bb] Ligand chirality error Hi all, How to check the presence of improper chirality of a fitted ligand. Is there any way to get rid of such error. All suggestions are welcome. Thanks in advance. sincerely Debajyoti http://sigads.rediff.com/RealMedia/ads/click_nx.ads/www.rediffmail.com/ signatureline.htm@Middle? Follow Rediff Deal ho jaye! http://track.rediff.com/click?url=___http://dealhojaye.rediff.com?sc_ci d=rediffmailsignature___cmp=signaturelnk=rediffmailsignaturenewservic e=deals to get exciting offers in your city everyday.
[ccp4bb] Postdoc opportunity at the Paterson Institute, Manchester
--- Postdoctoral position in DNA Damage Response Group at the Paterson Institute for Cancer Research, Manchester --- The Paterson Institute is a leading cancer centre of excellence core-funded by Cancer Research UK and is an Institute of The University of Manchester. A 3-years position with the possibility of extension is available in the DNA Damage Response group led by Dr Ivan Ahel to study structure and function of DNA repair enzymes: • Slade, D., Dunstan, M.S., Barkauskaite, E., Weston, R., Lafite, P., Dixon, N., Ahel, M., Leys, D., and Ahel, I. (2011). The structure and catalytic mechanism of a poly(ADP-ribose) glycohydrolase. Nature 477, 616–620. • Ahel, I., Ahel, D., Matsusaka, T., Clark, A.J., Pines, J., Boulton, S.J., and West, S.C. (2008). Poly(ADP-ribose)-binding zinc finger motifs in DNA repair/checkpoint proteins. Nature 451, 81–85. The successful candidate will have a PhD in biochemistry or cell biology, and at least one peer-reviewed publication as a first author. For information on the DNA Damage Response group, please visit: www.paterson.man.ac.uk/dnadamage/ Enquiries should be directed to Dr Ivan Ahel at: ia...@picr.man.ac.uk Closing date: 24th February 2012
Re: [ccp4bb] Kinase crystallization
It is a fairly common issue with kinases. Among other options you may want to try a generic kinase inhibitor (there are several good ones just look at pdb structures for ieas) and if this does not help then you could attempt to clamp the motion down via an inter-lobe engineered disulphide bond... Artem On Jan 30, 2012 5:17 AM, CHAVES SANJUAN, ANTONIO xanto...@iqfr.csic.es wrote: Dear all,** I am trying to crystallize a protein kinase without any success.** I suspect about its characteristic catalytic loop. I have already prepared different constructs, different expression vectors, and different mutant proteins (pseudo-phosphorylated, active, inactive?). I have also tested some ligands as ANPpnp (non hidrolizable nucleotide), ADP (product) and manganese (cofactor).** I am thinking in trying to cocrystallize it with a general kinase substrate (a peptide, a small molecule...). Does any one have any experience or suggestion?** Thanks in advance.** Sincerely,** Antonio
Re: [ccp4bb] Kinase crystallization
Staurosporine come to mind as a general kinase inhibitor. I also second Artem's suggestion that ligands make a big difference, we had several cases of kinases which required ligands for crystallization success. Also make sure you eliminate any floppy ends which may interfere with packing. Good luck Carsten From: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] On Behalf Of CHAVES SANJUAN, ANTONIO Sent: Monday, January 30, 2012 5:07 AM To: CCP4BB@JISCMAIL.AC.UK Subject: [ccp4bb] Kinase crystallization Dear all, I am trying to crystallize a protein kinase without any success. I suspect about its characteristic catalytic loop. I have already prepared different constructs, different expression vectors, and different mutant proteins (pseudo-phosphorylated, active, inactive?). I have also tested some ligands as ANPpnp (non hidrolizable nucleotide), ADP (product) and manganese (cofactor). I am thinking in trying to cocrystallize it with a general kinase substrate (a peptide, a small molecule...). Does any one have any experience or suggestion? Thanks in advance. Sincerely, Antonio
[ccp4bb] pointless eats FP column
Hi, I want determine the spacegroup with pointless and it should directly write out the mtz in the best sg. When I give a mtz with Structure Factor amplitude pointless recognize the file (9 columns) change the space group and the new file in the new space group has just 7 columns and FP and SigFP diappeared. Running pointless with the option match to reference it works fine and the all columns are there. I tried it from commandline and Interface but it does not work. I use pointless 1.6.5. Has anyone a suggestion what goes wrong here. Does it prevent user error when merged F's are given and omit them from the final mtz? However the input mtz was in P1 so it should be not a problem. Thanks in advance Best Regards Christian
Re: [ccp4bb] pointless eats FP column
Pointless can really only determine the space group from an unmerged file which wouldn't contain a merged amplitude F, so I'm not quite sure what you are trying to do Can you send me the file your command off-list I'll check best wishes Phil On 30 Jan 2012, at 13:54, Christian Roth wrote: Hi, I want determine the spacegroup with pointless and it should directly write out the mtz in the best sg. When I give a mtz with Structure Factor amplitude pointless recognize the file (9 columns) change the space group and the new file in the new space group has just 7 columns and FP and SigFP diappeared. Running pointless with the option match to reference it works fine and the all columns are there. I tried it from commandline and Interface but it does not work. I use pointless 1.6.5. Has anyone a suggestion what goes wrong here. Does it prevent user error when merged F's are given and omit them from the final mtz? However the input mtz was in P1 so it should be not a problem. Thanks in advance Best Regards Christian
Re: [ccp4bb] Reasoning for Rmeas or Rpim as Cutoff
Somebody sent this to me after a previous post a while back--a sort of case-study: Wang, J. (2010). Inclusion of weak high-resolution X-ray data for improvement of a group II intron structure. Acta crystallographica Section D, Biological crystallography 66, 988-1000. JPK On Mon, Jan 30, 2012 at 4:03 AM, Frank von Delft frank.vonde...@sgc.ox.ac.uk wrote: Hi Randy - thank you for a very interesting reminder to old literature. I'm intrigued: how come this apparently excellent idea has not become standard best practice in the 14 years since it was published? phx On 30/01/2012 09:40, Randy Read wrote: Hi, Here are a couple of links on the idea of judging resolution by a type of cross-validation with data not used in refinement: Ling et al, 1998: http://pubs.acs.org/doi/full/10.1021/bi971806n Brunger et al, 2008: http://journals.iucr.org/d/issues/2009/02/00/ba5131/index.html (cites earlier relevant papers from Brunger's group) Best wishes, Randy Read On 30 Jan 2012, at 07:09, arka chakraborty wrote: Hi all, In the context of the above going discussion can anybody post links for a few relevant articles? Thanks in advance, ARKO On Mon, Jan 30, 2012 at 3:05 AM, Randy Read rj...@cam.ac.uk wrote: Just one thing to add to that very detailed response from Ian. We've tended to use a slightly different approach to determining a sensible resolution cutoff, where we judge whether there's useful information in the highest resolution data by whether it agrees with calculated structure factors computed from a model that hasn't been refined against those data. We first did this with the complex of the Shiga-like toxin B-subunit pentamer with the Gb3 trisaccharide (Ling et al, 1998). From memory, the point where the average I/sig(I) drops below 2 was around 3.3A. However, we had a good molecular replacement model to solve this structure and, after just carrying out rigid-body refinement, we computed a SigmaA plot using data to the edge of the detector (somewhere around 2.7A, again from memory). The SigmaA plot dropped off smoothly to 2.8A resolution, with values well above zero (indicating significantly better than random agreement), then dropped suddenly. So we chose 2.8A as the cutoff. Because there were four pentamers in the asymmetric unit, we could then use 20-fold NCS averaging, which gave a fantastic map. In this case, the averaging certainly helped to pull out something very useful from a very weak signal, because the maps weren't nearly as clear at lower resolution. Since then, a number of other people have applied similar tests. Notably, Axel Brunger has done some careful analysis to show that it can indeed be useful to take data beyond the conventional limits. When you don't have a great MR model, you can do something similar by limiting the resolution for the initial refinement and rebuilding, then assessing whether there's useful information at higher resolution by using the improved model (which hasn't seen the higher resolution data) to compute Fcalcs. By the way, it's not necessary to use a SigmaA plot -- the correlation between Fo and Fc probably works just as well. Note that, when the model has been refined against the lower resolution data, you'll expect a drop in correlation at the resolution cutoff you used for refinement, unless you only use the cross-validation data for the resolution range used in refinement. - Randy J. Read Department of Haematology, University of Cambridge Cambridge Institute for Medical Research Tel: +44 1223 336500 Wellcome Trust/MRC Building Fax: +44 1223 336827 Hills Road E-mail: rj...@cam.ac.uk Cambridge CB2 0XY, U.K. www-structmed.cimr.cam.ac.uk On 29 Jan 2012, at 17:25, Ian Tickle wrote: Jacob, here's my (personal) take on this: The data quality metrics that everyone uses clearly fall into 2 classes: 'consistency' metrics, i.e. Rmerge/meas/pim and CC(1/2) which measure how well redundant observations agree, and signal/noise ratio metrics, i.e. mean(I/sigma) and completeness, which relate to the information content of the data. IMO the basic problem with all the consistency metrics is that they are not measuring the quantity that is relevant to refinement and electron density maps, namely the information content of the data, at least not in a direct and meaningful way. This is because there are 2 contributors to any consistency metric: the systematic errors (e.g. differences in illuminated volume and absorption) and the random errors (from counting statistics, detector noise etc.). If the data are collected with sufficient redundancy the systematic errors should hopefully largely cancel, and therefore only the random errors will determine the information content. Therefore the systematic error component of the consistency measure (which I suspect is the biggest component, at least for the strong
Re: [ccp4bb] pointless eats FP column
Dear Christian When I run this data into Pointless it doesn't find any higher symmetry than I422. However, the data appear to be highly twinned so I would be wary of believing that. If you put a merged file into Pointless, it can check for under-merging, but you can't really expect to use the output file as the program is now. It treats the file as if it was unmerged by squaring F to make a fake intensity, which is what is in the output file You can take the raw output from XDS into Pointless, either XDS_ASCII.HKL or INTEGRATE.HKL, and that will give you a better idea of the symmetry - you can then scale the output file with Scala or Aimless if you wish, though I have no reason to suppose that it is better than CORRECT XSCALE I'm copying this to the BB for general informaiton best wishes Phil On 30 Jan 2012, at 13:54, Christian Roth wrote: Hi, I want determine the spacegroup with pointless and it should directly write out the mtz in the best sg. When I give a mtz with Structure Factor amplitude pointless recognize the file (9 columns) change the space group and the new file in the new space group has just 7 columns and FP and SigFP diappeared. Running pointless with the option match to reference it works fine and the all columns are there. I tried it from commandline and Interface but it does not work. I use pointless 1.6.5. Has anyone a suggestion what goes wrong here. Does it prevent user error when merged F's are given and omit them from the final mtz? However the input mtz was in P1 so it should be not a problem. Thanks in advance Best Regards Christian
Re: [ccp4bb] Reasoning for Rmeas or Rpim as Cutoff
I'm intrigued: how come this apparently excellent idea has not become standard best practice in the 14 years since it was published? It would seem because too few people know about it, and it is not implemented in any software in the usual pipeline. Maybe it could be? Perhaps the way to do it would be always to integrate to ridiculously-high resolution, give that to Refmac, and starting from lower resolution, to iterate to higher resolution according the most recent sigma a calculation, and cutoff according to some reasonable sigma a value? JPK phx On 30/01/2012 09:40, Randy Read wrote: Hi, Here are a couple of links on the idea of judging resolution by a type of cross-validation with data not used in refinement: Ling et al, 1998: http://pubs.acs.org/doi/full/10.1021/bi971806n Brunger et al, 2008: http://journals.iucr.org/d/issues/2009/02/00/ba5131/index.html (cites earlier relevant papers from Brunger's group) Best wishes, Randy Read On 30 Jan 2012, at 07:09, arka chakraborty wrote: Hi all, In the context of the above going discussion can anybody post links for a few relevant articles? Thanks in advance, ARKO On Mon, Jan 30, 2012 at 3:05 AM, Randy Read rj...@cam.ac.uk wrote: Just one thing to add to that very detailed response from Ian. We've tended to use a slightly different approach to determining a sensible resolution cutoff, where we judge whether there's useful information in the highest resolution data by whether it agrees with calculated structure factors computed from a model that hasn't been refined against those data. We first did this with the complex of the Shiga-like toxin B-subunit pentamer with the Gb3 trisaccharide (Ling et al, 1998). From memory, the point where the average I/sig(I) drops below 2 was around 3.3A. However, we had a good molecular replacement model to solve this structure and, after just carrying out rigid-body refinement, we computed a SigmaA plot using data to the edge of the detector (somewhere around 2.7A, again from memory). The SigmaA plot dropped off smoothly to 2.8A resolution, with values well above zero (indicating significantly better than random agreement), then dropped suddenly. So we chose 2.8A as the cutoff. Because there were four pentamers in the asymmetric unit, we could then use 20-fold NCS averaging, which gave a fantastic map. In this case, the averaging certainly helped to pull out something very useful from a very weak signal, because the maps weren't nearly as clear at lower resolution. Since then, a number of other people have applied similar tests. Notably, Axel Brunger has done some careful analysis to show that it can indeed be useful to take data beyond the conventional limits. When you don't have a great MR model, you can do something similar by limiting the resolution for the initial refinement and rebuilding, then assessing whether there's useful information at higher resolution by using the improved model (which hasn't seen the higher resolution data) to compute Fcalcs. By the way, it's not necessary to use a SigmaA plot -- the correlation between Fo and Fc probably works just as well. Note that, when the model has been refined against the lower resolution data, you'll expect a drop in correlation at the resolution cutoff you used for refinement, unless you only use the cross-validation data for the resolution range used in refinement. - Randy J. Read Department of Haematology, University of Cambridge Cambridge Institute for Medical Research Tel: +44 1223 336500 Wellcome Trust/MRC Building Fax: +44 1223 336827 Hills Road E-mail: rj...@cam.ac.uk Cambridge CB2 0XY, U.K. www-structmed.cimr.cam.ac.uk On 29 Jan 2012, at 17:25, Ian Tickle wrote: Jacob, here's my (personal) take on this: The data quality metrics that everyone uses clearly fall into 2 classes: 'consistency' metrics, i.e. Rmerge/meas/pim and CC(1/2) which measure how well redundant observations agree, and signal/noise ratio metrics, i.e. mean(I/sigma) and completeness, which relate to the information content of the data. IMO the basic problem with all the consistency metrics is that they are not measuring the quantity that is relevant to refinement and electron density maps, namely the information content of the data, at least not in a direct and meaningful way. This is because there are 2 contributors to any consistency metric: the systematic errors (e.g. differences in illuminated volume and absorption) and the random errors (from counting statistics, detector noise etc.). If the data are collected with sufficient redundancy the systematic errors should hopefully largely cancel, and therefore only the random errors will determine the information content. Therefore the systematic error component of the consistency measure (which I suspect is the biggest component, at least for the strong reflections) is not
Re: [ccp4bb] Reasoning for Rmeas or Rpim as Cutoff
On Jan 30, 2012, at 10:28 AM, Jacob Keller wrote: I'm intrigued: how come this apparently excellent idea has not become standard best practice in the 14 years since it was published? It would seem because too few people know about it, and it is not implemented in any software in the usual pipeline. Maybe it could be? Phenix.model_vs_data calculates a sigmaA_ vs resolution plot (in comprehensive validation in the GUI). Pavel would probably have replied by now, but I don't think the discussion has been cross-posted to the phenix bb. Cheers, Florian Perhaps the way to do it would be always to integrate to ridiculously-high resolution, give that to Refmac, and starting from lower resolution, to iterate to higher resolution according the most recent sigma a calculation, and cutoff according to some reasonable sigma a value? JPK phx On 30/01/2012 09:40, Randy Read wrote: Hi, Here are a couple of links on the idea of judging resolution by a type of cross-validation with data not used in refinement: Ling et al, 1998: http://pubs.acs.org/doi/full/10.1021/bi971806n Brunger et al, 2008: http://journals.iucr.org/d/issues/2009/02/00/ba5131/index.html (cites earlier relevant papers from Brunger's group) Best wishes, Randy Read On 30 Jan 2012, at 07:09, arka chakraborty wrote: Hi all, In the context of the above going discussion can anybody post links for a few relevant articles? Thanks in advance, ARKO On Mon, Jan 30, 2012 at 3:05 AM, Randy Read rj...@cam.ac.uk wrote: Just one thing to add to that very detailed response from Ian. We've tended to use a slightly different approach to determining a sensible resolution cutoff, where we judge whether there's useful information in the highest resolution data by whether it agrees with calculated structure factors computed from a model that hasn't been refined against those data. We first did this with the complex of the Shiga-like toxin B-subunit pentamer with the Gb3 trisaccharide (Ling et al, 1998). From memory, the point where the average I/sig(I) drops below 2 was around 3.3A. However, we had a good molecular replacement model to solve this structure and, after just carrying out rigid-body refinement, we computed a SigmaA plot using data to the edge of the detector (somewhere around 2.7A, again from memory). The SigmaA plot dropped off smoothly to 2.8A resolution, with values well above zero (indicating significantly better than random agreement), then dropped suddenly. So we chose 2.8A as the cutoff. Because there were four pentamers in the asymmetric unit, we could then use 20-fold NCS averaging, which gave a fantastic map. In this case, the averaging certainly helped to pull out something very useful from a very weak signal, because the maps weren't nearly as clear at lower resolution. Since then, a number of other people have applied similar tests. Notably, Axel Brunger has done some careful analysis to show that it can indeed be useful to take data beyond the conventional limits. When you don't have a great MR model, you can do something similar by limiting the resolution for the initial refinement and rebuilding, then assessing whether there's useful information at higher resolution by using the improved model (which hasn't seen the higher resolution data) to compute Fcalcs. By the way, it's not necessary to use a SigmaA plot -- the correlation between Fo and Fc probably works just as well. Note that, when the model has been refined against the lower resolution data, you'll expect a drop in correlation at the resolution cutoff you used for refinement, unless you only use the cross-validation data for the resolution range used in refinement. - Randy J. Read Department of Haematology, University of Cambridge Cambridge Institute for Medical ResearchTel: +44 1223 336500 Wellcome Trust/MRC Building Fax: +44 1223 336827 Hills Road E-mail: rj...@cam.ac.uk Cambridge CB2 0XY, U.K. www-structmed.cimr.cam.ac.uk On 29 Jan 2012, at 17:25, Ian Tickle wrote: Jacob, here's my (personal) take on this: The data quality metrics that everyone uses clearly fall into 2 classes: 'consistency' metrics, i.e. Rmerge/meas/pim and CC(1/2) which measure how well redundant observations agree, and signal/noise ratio metrics, i.e. mean(I/sigma) and completeness, which relate to the information content of the data. IMO the basic problem with all the consistency metrics is that they are not measuring the quantity that is relevant to refinement and electron density maps, namely the information content of the data, at least not in a direct and meaningful way. This is because there are 2 contributors to any consistency metric: the systematic errors (e.g. differences in illuminated volume and absorption) and the random errors (from counting statistics, detector noise etc.). If the data are collected with
Re: [ccp4bb] B_sol from EDS
Hi Bernhard, I just calculated k_sol and B_sol for all PDB entries that - have reflection data available, - I could re-compute the R-factor within 5%, and - R-work30% using a simple cctbx script. Here is what I get: Distribution of k_sol: 0.000 - 0.060 : 27 0.060 - 0.120 : 12 0.120 - 0.180 : 51 0.180 - 0.240 : 182 0.240 - 0.300 : 1770 0.300 - 0.360 : 13819 0.360 - 0.420 : 19731 0.420 - 0.480 : 3039 0.480 - 0.540 : 471 0.540 - 0.600 : 256 Distribution of B_sol: 0.000 - 31.300 : 4349 31.300 - 62.600 : 29425 62.600 - 93.900 : 4578 93.900 - 125.200: 597 125.200 - 156.500: 225 156.500 - 187.800: 84 187.800 - 219.100: 37 219.100 - 250.400: 23 250.400 - 281.700: 10 281.700 - 313.000: 30 It seems like the result of similar exercise done by Fokine and Urzhumtsev (Acta Cryst. (2002). D58, 1387-1392) still holds (see figure 3 on page 1390 there). Pavel On Mon, Jan 30, 2012 at 11:10 AM, Bernhard Rupp (Hofkristallrat a.D.) hofkristall...@gmail.com wrote: Dear All, when I plot bulk solvent B and K extracted from EDS, an improbable and bimodal distribution appears. In the B_sol vs k_sol PDF a sharp line of values with B-sol of 70 appears (B-axis left to right, 0-200). http://www.ruppweb.org/images/b_sol_contour.jpg http://www.ruppweb.org/images/b_sol_surface.jpg According to a quick peak at EDS instructions, it uses the REFMAC flat bulk solvent model throughout for bulk solvent correction. The main peak in fact has the expected distribution, but it seems that the sharp peak at B_sol=70 represents some cut-off that in a certain set of calculations was used. For data mining it would be useful to know where/when these cutoffs were used. Best regards, BR - Bernhard Rupp http://www.ruppweb.org/ -
Re: [ccp4bb] Reasoning for Rmeas or Rpim as Cutoff
Frank, Don't you already get a plot of SigmaA versus resolution from refmac, where the free set of reflections has been used to estimate SigmaA? Have a look at some of your log files. Pete From: CCP4 bulletin board [CCP4BB@JISCMAIL.AC.UK] On Behalf Of Frank von Delft [frank.vonde...@sgc.ox.ac.uk] Sent: Monday, January 30, 2012 2:03 AM To: CCP4BB@JISCMAIL.AC.UK Subject: Re: [ccp4bb] Reasoning for Rmeas or Rpim as Cutoff Hi Randy - thank you for a very interesting reminder to old literature. I'm intrigued: how come this apparently excellent idea has not become standard best practice in the 14 years since it was published? phx On 30/01/2012 09:40, Randy Read wrote: Hi, Here are a couple of links on the idea of judging resolution by a type of cross-validation with data not used in refinement: Ling et al, 1998: http://pubs.acs.org/doi/full/10.1021/bi971806n Brunger et al, 2008: http://journals.iucr.org/d/issues/2009/02/00/ba5131/index.html (cites earlier relevant papers from Brunger's group) Best wishes, Randy Read On 30 Jan 2012, at 07:09, arka chakraborty wrote: Hi all, In the context of the above going discussion can anybody post links for a few relevant articles? Thanks in advance, ARKO On Mon, Jan 30, 2012 at 3:05 AM, Randy Read rj...@cam.ac.ukmailto:rj...@cam.ac.uk wrote: Just one thing to add to that very detailed response from Ian. We've tended to use a slightly different approach to determining a sensible resolution cutoff, where we judge whether there's useful information in the highest resolution data by whether it agrees with calculated structure factors computed from a model that hasn't been refined against those data. We first did this with the complex of the Shiga-like toxin B-subunit pentamer with the Gb3 trisaccharide (Ling et al, 1998). From memory, the point where the average I/sig(I) drops below 2 was around 3.3A. However, we had a good molecular replacement model to solve this structure and, after just carrying out rigid-body refinement, we computed a SigmaA plot using data to the edge of the detector (somewhere around 2.7A, again from memory). The SigmaA plot dropped off smoothly to 2.8A resolution, with values well above zero (indicating significantly better than random agreement), then dropped suddenly. So we chose 2.8A as the cutoff. Because there were four pentamers in the asymmetric unit, we could then use 20-fold NCS averaging, which gave a fantastic map. In this case, the averaging certainly helped to pull out something very useful from a very weak signal, because the maps weren't nearly as clear at lower resolution. Since then, a number of other people have applied similar tests. Notably, Axel Brunger has done some careful analysis to show that it can indeed be useful to take data beyond the conventional limits. When you don't have a great MR model, you can do something similar by limiting the resolution for the initial refinement and rebuilding, then assessing whether there's useful information at higher resolution by using the improved model (which hasn't seen the higher resolution data) to compute Fcalcs. By the way, it's not necessary to use a SigmaA plot -- the correlation between Fo and Fc probably works just as well. Note that, when the model has been refined against the lower resolution data, you'll expect a drop in correlation at the resolution cutoff you used for refinement, unless you only use the cross-validation data for the resolution range used in refinement. - Randy J. Read Department of Haematology, University of Cambridge Cambridge Institute for Medical ResearchTel: +44 1223 336500tel:%2B44%201223%20336500 Wellcome Trust/MRC Building Fax: +44 1223 336827tel:%2B44%201223%20336827 Hills RoadE-mail: rj...@cam.ac.ukmailto:rj...@cam.ac.uk Cambridge CB2 0XY, U.K. www-structmed.cimr.cam.ac.ukhttp://www-structmed.cimr.cam.ac.uk/ On 29 Jan 2012, at 17:25, Ian Tickle wrote: Jacob, here's my (personal) take on this: The data quality metrics that everyone uses clearly fall into 2 classes: 'consistency' metrics, i.e. Rmerge/meas/pim and CC(1/2) which measure how well redundant observations agree, and signal/noise ratio metrics, i.e. mean(I/sigma) and completeness, which relate to the information content of the data. IMO the basic problem with all the consistency metrics is that they are not measuring the quantity that is relevant to refinement and electron density maps, namely the information content of the data, at least not in a direct and meaningful way. This is because there are 2 contributors to any consistency metric: the systematic errors (e.g. differences in illuminated volume and absorption) and the random errors (from counting statistics, detector noise etc.). If the data are collected with sufficient redundancy the
[ccp4bb] Staff Scientist position at MacCHESS
The Macromolecular Diffraction Facility of the Cornell High-Energy Synchrotron Source (MacCHESS) has an opening for a Staff Scientist (Research Associate) to pursue the development of novel techniques in x-ray scattering as applied to structural biology, and to support users at MacCHESS. There will also be opportunities to pursue projects in structural biology, using current crystallographic and SAXS methods. Research areas of particular interest include structure solution from multiple crystals, use of Laue diffraction, BioSAXS, microfluidics, and user interfaces for beamline operation. A Ph.D. in structural biology, biophysics, or a related field, and at least 3 years of experience beyond the degree in a relevant field is required. A solid publication record is essential, and experience working at a synchrotron facility is highly desirable. Excellent communication skills are a must, including fluency in the English language. Appointments are nominally for three years with the possibility for renewal, subject to mutual satisfaction and the availability of funds. Located on an Ivy League university campus in picturesque upstate New York, the Cornell High-Energy Synchrotron Source (CHESS) serves a world-wide user base of structural biologists, chemists, physicists, and engineers. MacCHESS is an NIH-supported National Resource providing support for structural biology at CHESS. MacCHESS is a heavily team-oriented environment. Please provide an application and have at least three letters of reference sent to: Dr. Marian Szebenyi, Chair MacCHESS Staff Scientist Search Committee c/o Peggy Steenrod Newman Laboratory Cornell University Ithaca, NY 14853 USA Applications should include a cover letter, curriculum vita, a publication list, and a detailed summary of research experience and interests. Electronic submissions and inquiries may be addressed to search-cla...@cornell.edu. Salary and starting date are negotiable. Cornell is an equal opportunity, affirmative action educator and employer.
Re: [ccp4bb] B_sol from EDS
Yes, that is about what one would expect. I also checked a few of the extreme outliers, and almost always can come up with a reasonable value. Which does not remove my curiosity regarding the B_sol 70 cutoff and its purpose. Cheers, BR From: Pavel Afonine [mailto:pafon...@gmail.com] Sent: Monday, January 30, 2012 11:33 AM To: b...@hofkristallamt.org Cc: CCP4BB@jiscmail.ac.uk Subject: Re: [ccp4bb] B_sol from EDS Hi Bernhard, I just calculated k_sol and B_sol for all PDB entries that - have reflection data available, - I could re-compute the R-factor within 5%, and - R-work30% using a simple cctbx script. Here is what I get: Distribution of k_sol: 0.000 - 0.060 : 27 0.060 - 0.120 : 12 0.120 - 0.180 : 51 0.180 - 0.240 : 182 0.240 - 0.300 : 1770 0.300 - 0.360 : 13819 0.360 - 0.420 : 19731 0.420 - 0.480 : 3039 0.480 - 0.540 : 471 0.540 - 0.600 : 256 Distribution of B_sol: 0.000 - 31.300 : 4349 31.300 - 62.600 : 29425 62.600 - 93.900 : 4578 93.900 - 125.200: 597 125.200 - 156.500: 225 156.500 - 187.800: 84 187.800 - 219.100: 37 219.100 - 250.400: 23 250.400 - 281.700: 10 281.700 - 313.000: 30 It seems like the result of similar exercise done by Fokine and Urzhumtsev (Acta Cryst. (2002). D58, 1387-1392) still holds (see figure 3 on page 1390 there). Pavel On Mon, Jan 30, 2012 at 11:10 AM, Bernhard Rupp (Hofkristallrat a.D.) hofkristall...@gmail.com wrote: Dear All, when I plot bulk solvent B and K extracted from EDS, an improbable and bimodal distribution appears. In the B_sol vs k_sol PDF a sharp line of values with B-sol of 70 appears (B-axis left to right, 0-200). http://www.ruppweb.org/images/b_sol_contour.jpg http://www.ruppweb.org/images/b_sol_surface.jpg According to a quick peak at EDS instructions, it uses the REFMAC flat bulk solvent model throughout for bulk solvent correction. The main peak in fact has the expected distribution, but it seems that the sharp peak at B_sol=70 represents some cut-off that in a certain set of calculations was used. For data mining it would be useful to know where/when these cutoffs were used. Best regards, BR - Bernhard Rupp http://www.ruppweb.org/ -
[ccp4bb] odd behaviour of reindex
Hi all, we encountered an odd behaviour of REINDEX. Snip form logfile: Data line--- reindex HKL (h+l)/2, -k, (h-l)/2 Data line--- end Reflections will be reindexed, and unit cell recalculated Reindexing transformation: (h' k' l') = ( h k l ) ( 1.0 0.0 1.0 ) ( 0.0 -1.0 0.0 ) ( 0.5 0.0 -0.5 ) Obviously, the first line of the matrix is not what we intended to create. inputting the transformation as HKL h/2+l/2, -k, h/2-l/2 produces the desired result: Data line--- reindex HKL h/2+l/2, -k, h/2-l/2 Data line--- end Reflections will be reindexed, and unit cell recalculated Reindexing transformation: (h' k' l') = ( h k l ) ( 0.5 0.0 0.5 ) ( 0.0 -1.0 0.0 ) ( 0.5 0.0 -0.5 ) Admittedly, the documentation does not use any brackets in the examples, but i would expect REINDEX either to throw an error or treat (h+l)/2 like (h-l)/2 but not treat them in the way encountered. Cheers, Jens -- +-+-+ | Jens T. Kaiser | Office: +1(626)395-2662 | | California Institute of Technology | Lab:+1(626)395-8392 | | m/c 114-96 | Cell: +1(626)379-1650 | | 1200 E. California Blvd.| Xray: +1(626)395-2661 | | Pasadena, CA 91125 | Email: kai...@caltech.edu | | USA | Skype: jens.t.kaiser | +-+-+
Re: [ccp4bb] Kinase crystallization
if the ligand binding site is exposed to the solvent a bound ligand may help. if the protein has is flexible domains and ligand fix it in one conformation, a bound ligand will help even more. All the above assuming that the purity and the concentration of the protein are high. George From: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] On Behalf Of Debreczeni, Judit Sent: Monday, January 30, 2012 3:45 PM To: CCP4BB@JISCMAIL.AC.UK Subject: Re: [ccp4bb] Kinase crystallization Does your kinase autophosphorylate by any chance? -- That can produce differently phosphorylated species and affect crystallisability. You can detect it by e.g. mass spec, and tackle it by dephosphorylating the protein prior to crystallisation or by coexpression with a phosphatase. _ AstraZeneca UK Limited is a company incorporated in England and Wales with registered number: 03674842 and a registered office at 2 Kingdom Street, London, W2 6BD. Confidentiality Notice: This message is private and may contain confidential, proprietary and legally privileged information. If you have received this message in error, please notify us and remove it from your system and note that you must not copy, distribute or take any action in reliance on it. Any unauthorised use or disclosure of the contents of this message is not permitted and may be unlawful. Disclaimer: Email messages may be subject to delays, interception, non-delivery and unauthorised alterations. Therefore, information expressed in this message is not given or endorsed by AstraZeneca UK Limited unless otherwise notified by an authorised representative independent of this message. No contractual relationship is created by this message by any person unless specifically indicated by agreement in writing other than email. Monitoring: AstraZeneca UK Limited may monitor email traffic data and content for the purposes of the prevention and detection of crime, ensuring the security of our computer systems and checking compliance with our Code of Conduct and policies. From: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] On Behalf Of CHAVES SANJUAN, ANTONIO Sent: 30 January 2012 10:07 To: CCP4BB@JISCMAIL.AC.UK Subject: [ccp4bb] Kinase crystallization Dear all, I am trying to crystallize a protein kinase without any success. I suspect about its characteristic catalytic loop. I have already prepared different constructs, different expression vectors, and different mutant proteins (pseudo-phosphorylated, active, inactive?). I have also tested some ligands as ANPpnp (non hidrolizable nucleotide), ADP (product) and manganese (cofactor). I am thinking in trying to cocrystallize it with a general kinase substrate (a peptide, a small molecule...). Does any one have any experience or suggestion? Thanks in advance. Sincerely, Antonio
