Re: [ccp4bb] database-assisted data archive
The PiMS team intends that the CCP4 records link not only with the synchrotron, but further back to crystallogenesis records in xtalPiMS, and protein production records in PiMS. The benefits this will provide include: - if you find an unexpected piece of electron density, navigating to records that show what substances were in the sample - designing crystalogenesis screens in the light of data not only about crystals obtained, but also about diffraction. Paul Paukstelis rightly points out that was convincing anyone to actually use it is hard, even though the cost of lost work is significant. To address this, we need to ensure: - data entry is as automatic as possible - everything joins up, so that one act of data entry has multiple payoffs. The aim must be seamless data transfer and consistent user interfaces, all the way from target selection to structure interpretation, delivered in a way that is extensible as methods evolve, and which supports not only PX but also other methods. This is a large challenge, but it is achievable. Andreas, in the short term I suggest you look at keeping your files in a Subversion repository. This provides a central backup, and it can easily be mapped as a folder on Linux, OSX, and Windows, because it implements the WebDAV standard. Each project can have a sub-folder. regards, Chris Chris Morris chris.mor...@stfc.ac.uk Tel: +44 (0)1925 603689 Fax: +44 (0)1925 603634 Mobile: 07921-717915 https://www.pims-lims.org/ Daresbury Lab, Daresbury, Warrington, UK, WA4 4AD Date:Wed, 18 Aug 2010 12:19:36 +0100 From:Georgios Pelios george.pel...@stfc.ac.uk Subject: Re: database-assisted data archive Dear all As CCP4, we are currently developing the new CCP4i that will include a database application that will store project and job data. The database schema has already been designed but its design is not final and can be modified depending on user feedback. Now, we are in the process of writing the database API. Any suggestions and ideas regarding data storage and retrieval are welcome. George Pelios CCP4
[ccp4bb] and another MR problem
Dear all I have a molecular replacement problem and I really need your help with this one. My protein crystallized in C2 instead of P21 like it did before. Now when I tried to do MR to place it in C2 it doesn't work, I get no solution. This protein has been crystallized in P212121, P31, P43 and C2 and it has two loops in different regions that depending on the packing, they'll show density so it is flexible. When I crystallized it the first time in P21 and used a publish model I had no problem at all to place it, just right now. I've have tried molrep, phaser and phenix with different resolution ranges (40-3.0, 20-3.0, 10-3.0) and different models, i.e. publish structures or the one I got at P21. I also tried doing rotation first at different resolutions then compare the peaks obtained and picked the ones that are consistent and give that for translation search and yet no solution. The resolution is 2.4 A and the Matthews coefficient predicts 3 molecules in ASU with 43.6% solvent. I also tried searching for 2 molecules just in case that it is high solvent content (62.5%) as monomer and as dimer and yet no solution. The data itself processed well and I've processed in HKL and d*trek and used both processed files and yet no solution. Completeness is 96.9% overall and 82% (2.49-2.40) shell. Redundancy is 3.5 overall and 2.6 at the(2.49-2.40) shell. Here is the table from HKL: Lower Upper Average Average Norm. Linear Square Angstrom I errorstat. Chi**2 R-fac R-fac 50.00 5.17 901.935.910.2 1.105 0.041 0.051 5.17 4.10 962.847.711.6 0.971 0.045 0.049 4.10 3.58 496.425.9 9.2 1.194 0.063 0.067 3.58 3.26 309.915.2 8.4 1.360 0.069 0.067 3.26 3.02 153.9 9.9 7.2 1.456 0.105 0.098 3.02 2.85 100.8 7.3 6.9 1.864 0.141 0.120 2.85 2.7057.4 7.0 6.8 1.584 0.200 0.162 2.70 2.5941.9 7.5 7.0 1.348 0.246 0.220 2.59 2.4930.5 7.4 7.3 1.526 0.329 0.285 2.49 2.4022.5 8.0 7.8 1.347 0.388 0.361 All reflections319.917.6 8.3 1.366 0.065 0.056 Would you please, please give me some tips, tricks, advice, encouragement? :-) Thank you so much Teresa Teresa De la Mora-Rey Ph.D. Dept. Medicinal Chemistry University of Minnesota 8-101 Weaver-Densford Hall 308 Harvard St. SE, Minneapolis, MN 55455 Lab phone (612) 626-5226 If you never did you should. These things are fun and fun is good Dr. Seuss
Re: [ccp4bb] and another MR problem
When doing MR, I usually try Phaser and EPMR. Phaser rarely fails for a high-homology MR search, but can have difficulty fitting 3 or more protein units in the ASU. EPMR is a little better, in my experience, in ferreting out a solution for 3 or more protein units. If you are running the Matthews Probability Calculator in Phaser (I think it is called Cell Content Analysis) you should take that result with a grain of salt, and consider additional solutions, e.g., 2 and 4 chains if the suggestion is 3 chains per ASU. While most proteins fall within the expected range of solvent content, it is possible to have solutions on the fringes. (We just had one recently with 67% solvent content.) In general, searching with larger protein units is better than with monomers, if you know the structure of the multimers. That is, searching with dimers instead of monomers is usually more effective, especially if this will reduce the number of protein units to be placed to 3 or fewer. If the biological unit is a tetramer, sometimes you can get a good solution with AB dimers but not with BC, CD, or AD dimers, etc. If running Phaser, be sure to allow for extra clashes in your search, or you may reject all of the possible solutions. It is not unreasonable to allow for 30 or more clashes in your initial trials. Also, you may want to retain 65% the best rotation peaks instead of the default 75%, to improve your chances of finding a solution. If Phaser doesn't work, try EPMR. I have not had it fail yet for an MR search for 3 or fewer protein chains per ASU, and high-homology search models. The success of EPMR can be improved by including a little more of the high-resolution data, although this will slow things down. In a difficult case, we extended the data used by EPMR almost to the diffraction limit of the crystal to get a good MR solution. The good news is that C2 is a relatively simple search space: no alternative space groups or screw axis combinations, or reindexing of data required. If you suspect 2 or 3 protein chains in the ASU, try for a partial solution for 1 or 2 chains, then examine the result for packing, which often gives you some clues as to where and how many additional chains might be placed. Cheers, and good luck. On 8/23/2010 12:10 PM, Teresa De la Mora wrote: Dear all I have a molecular replacement problem and I really need your help with this one. My protein crystallized in C2 instead of P21 like it did before. Now when I tried to do MR to place it in C2 it doesn't work, I get no solution. This protein has been crystallized in P212121, P31, P43 and C2 and it has two loops in different regions that depending on the packing, they'll show density so it is flexible. When I crystallized it the first time in P21 and used a publish model I had no problem at all to place it, just right now. I've have tried molrep, phaser and phenix with different resolution ranges (40-3.0, 20-3.0, 10-3.0) and different models, i.e. publish structures or the one I got at P21. I also tried doing rotation first at different resolutions then compare the peaks obtained and picked the ones that are consistent and give that for translation search and yet no solution. The resolution is 2.4 A and the Matthews coefficient predicts 3 molecules in ASU with 43.6% solvent. I also tried searching for 2 molecules just in case that it is high solvent content (62.5%) as monomer and as dimer and yet no solution. The data itself processed well and I've processed in HKL and d*trek and used both processed files and yet no solution. Completeness is 96.9% overall and 82% (2.49-2.40) shell. Redundancy is 3.5 overall and 2.6 at the(2.49-2.40) shell.Here is the table from HKL: Lower Upper Average Average Norm. Linear Square Angstrom Ierror stat. Chi**2 R-fac R-fac 50.00 5.17 901.9 35.9 10.2 1.105 0.041 0.051 5.17 4.10 962.8 47.7 11.6 0.971 0.045 0.049 4.10 3.58 496.4 25.9 9.2 1.194 0.063 0.067 3.58 3.26 309.9 15.2 8.4 1.360 0.069 0.067 3.26 3.02 153.9 9.9 7.2 1.456 0.105 0.098 3.02 2.85 100.8 7.3 6.9 1.864 0.141 0.120 2.85 2.70 57.47.0 6.8 1.584 0.200 0.162 2.70 2.59 41.97.5 7.0 1.348 0.246 0.220 2.59 2.49
Re: [ccp4bb] and another MR problem
Dear all The P21 unit cell is a=86.499 b=38.554 c=94.027 al=90.0 be=97.92 ga=90.0 The C2 unit cell is a=103.932 b=62.253 c=96.305 al=90.0 be=111.934 ga=90.0 Phaser and the other programs do give me solutions but the LLG in Phaser is about -1000's and the highest TFZ score I get is 4.9, and for molrep the Rfac is about 0.600's. I even tried to do rigid body with the solutions obtained and I get an increase value for Rfact and Rfree of 55-56%. Thank you all for your suggestions. I'll use them right now. Teresa On Aug 23, 2010, at 11:43 AM, Peter Zwart wrote: can you list you cell constants? On 23 August 2010 09:10, Teresa De la Mora dela0...@umn.edu wrote: Dear all I have a molecular replacement problem and I really need your help with this one. My protein crystallized in C2 instead of P21 like it did before. Now when I tried to do MR to place it in C2 it doesn't work, I get no solution. This protein has been crystallized in P212121, P31, P43 and C2 and it has two loops in different regions that depending on the packing, they'll show density so it is flexible. When I crystallized it the first time in P21 and used a publish model I had no problem at all to place it, just right now. I've have tried molrep, phaser and phenix with different resolution ranges (40-3.0, 20-3.0, 10-3.0) and different models, i.e. publish structures or the one I got at P21. I also tried doing rotation first at different resolutions then compare the peaks obtained and picked the ones that are consistent and give that for translation search and yet no solution. The resolution is 2.4 A and the Matthews coefficient predicts 3 molecules in ASU with 43.6% solvent. I also tried searching for 2 molecules just in case that it is high solvent content (62.5%) as monomer and as dimer and yet no solution. The data itself processed well and I've processed in HKL and d*trek and used both processed files and yet no solution. Completeness is 96.9% overall and 82% (2.49-2.40) shell. Redundancy is 3.5 overall and 2.6 at the(2.49-2.40) shell. Here is the table from HKL: Lower Upper Average Average Norm. Linear Square Angstrom I errorstat. Chi**2 R-fac R- fac 50.00 5.17 901.935.910.2 1.105 0.041 0.051 5.17 4.10 962.847.711.6 0.971 0.045 0.049 4.10 3.58 496.425.9 9.2 1.194 0.063 0.067 3.58 3.26 309.915.2 8.4 1.360 0.069 0.067 3.26 3.02 153.9 9.9 7.2 1.456 0.105 0.098 3.02 2.85 100.8 7.3 6.9 1.864 0.141 0.120 2.85 2.7057.4 7.0 6.8 1.584 0.200 0.162 2.70 2.5941.9 7.5 7.0 1.348 0.246 0.220 2.59 2.4930.5 7.4 7.3 1.526 0.329 0.285 2.49 2.4022.5 8.0 7.8 1.347 0.388 0.361 All reflections319.917.6 8.3 1.366 0.065 0.056 Would you please, please give me some tips, tricks, advice, encouragement? :-) Thank you so much Teresa Teresa De la Mora-Rey Ph.D. Dept. Medicinal Chemistry University of Minnesota 8-101 Weaver-Densford Hall 308 Harvard St. SE, Minneapolis, MN 55455 Lab phone (612) 626-5226 If you never did you should. These things are fun and fun is good Dr. Seuss -- - P.H. Zwart Research Scientist Berkeley Center for Structural Biology Lawrence Berkeley National Laboratories 1 Cyclotron Road, Berkeley, CA-94703, USA Cell: 510 289 9246 BCSB: http://bcsb.als.lbl.gov PHENIX: http://www.phenix-online.org SASTBX: http://sastbx.als.lbl.gov - Teresa De la Mora-Rey Ph.D. Dept. Medicinal Chemistry University of Minnesota 8-101 Weaver-Densford Hall 308 Harvard St. SE, Minneapolis, MN 55455 Lab phone (612) 626-5226 If you never did you should. These things are fun and fun is good Dr. Seuss
Re: [ccp4bb] Disulfide Designer Program
Hello, You could try Disulfide by Design: http://cptweb.cpt.wayne.edu/DbD/ Good luck, Konstantin On Mon, 23 Aug 2010, Jacob Keller wrote: Dear Crystallographers, I remember having heard of a program which takes a given oligomeric assembly, and suggests optimum disulfides to stablize the complex. Can someone refresh my memory which program that is, and where it is available? Best Regards, Jacob Keller *** 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 *** -- Konstantin Korotkov, Ph.D. Research Scientist University of Washington Department of Biochemistry Box 357742 Seattle, WA 98195-7742 (206)616-4512 k...@u.washington.edu --
Re: [ccp4bb] Disulfide Designer Program
I think that Disulfide by Design, or DbD, is geared toward monomeric proteins (ie. for enhancing stability). I'm sure you could get it to work with an oligomer with a little tinkering, but there is another server called sGAL that may be more like what you're looking for: http://bioinformatics.oxfordjournals.org/cgi/content/abstract/22/24/3101 Also, I have designed intermolecular disulfide bonds using the following method (although it is a bit of a hacked up way to do things, but it worked well in my situation). Put your PDB files for each chain of the oligomer through the following diffusion accessibility server: http://nihserver.mbi.ucla.edu/diff_acc/ This server will take your PDB file and rewrite the B-factor column with values for the diffusion accessibility of each atom. Now you can look at your PDB files in PyMol, and if you choose color by B-factor you will actually be coloring the atoms by diffusion accessibility. Do this for each chain of the oligomer separately. Now take those individual chains, colored by diff. accessibility and overlay them onto the structure of the oligomer. Look for residues whose C(gamma) are both accessible at the surface of each chain, and are within reasonable distance for a disulfide bond. You can check this by just measuring the distance, or if you want to be more precise, you can take several ideal disulfide bonds from other structures and compare the distances (and dihedral angles) by overlaying them onto your selected residues. If you find a good set of residues that match all the criteria I listed, chances are they will be good candidates to disulfide bond with each other if mutated to Cys. Remember that with disulfide bonds, geometry (dihedral angles) is very important in addition to the bond length. I don't recall off the top of my head what values for the dihedral angle and bond length are ideal, but this info is readily available in the literature. And also, your bond will never form if the gamma position of the side chain is buried by surrounding residues (thus, the use of the diff. accessibility server - I learned this the hard way by making many unsuccessful mutants). Again, this is kind of a drawn out way of doing things compared to just throwing your PDB file at a server like sGAL, but I feel that it is a bit more rigorous, and it has been successful in my hands. Good Luck, Mike Thompson - Original Message - From: Konstantin v. Korotkov k...@u.washington.edu To: CCP4BB@JISCMAIL.AC.UK Sent: Monday, August 23, 2010 4:09:50 PM GMT -08:00 US/Canada Pacific Subject: Re: [ccp4bb] Disulfide Designer Program Hello, You could try Disulfide by Design: http://cptweb.cpt.wayne.edu/DbD/ Good luck, Konstantin On Mon, 23 Aug 2010, Jacob Keller wrote: Dear Crystallographers, I remember having heard of a program which takes a given oligomeric assembly, and suggests optimum disulfides to stablize the complex. Can someone refresh my memory which program that is, and where it is available? Best Regards, Jacob Keller *** 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 *** -- Konstantin Korotkov, Ph.D. Research Scientist University of Washington Department of Biochemistry Box 357742 Seattle, WA 98195-7742 (206)616-4512 k...@u.washington.edu -- -- Michael C. Thompson Graduate Student Biochemistry Molecular Biology Division Department of Chemistry Biochemistry University of California, Los Angeles mi...@chem.ucla.edu