Re: [ccp4bb] coot problems to decrease R FREE
Hi Robbie, I agree that you bias R-free after the real-space refinement well, ok, isn't it enough to realize that this is bad and should be avoided ? (I guess we all know we should never bias Rfree!) My point was that we normally do not calculate R-free after real-space refinement, It's not about whether you compute something or not. It's about whether you expose free-r reflections to refinement and to current model, given that they should never ever see any refinement or model or optimization under any circumstances. Otherwise they immediately become non-free. but after reciprocal space refinement. Here the bias is removed again. Hm.. How you know this? I guess it requires a great deal of effort to prove this!. In practical terms we have to choose the best option: 1) Refine against maps with missing reflections and the possibility of artefacts that lead to suboptimal results. This keeps R-free unbiased, even directly after real-space refinement. 2) Refine against maps with test reflections set to Fcalc. This biases against the current model, but should have less artefacts. This too keeps R-free unbiased directly after real-space refinement. 3) Refine against maps with all reflections. This should give the best fitting results, but does introduce bias to the test set. However, this bias These are all valid points. However they do not advocate for biasing Rfree. Yuo can always fill in missing reflections with something else (different from your genuine free-r set of reelections) and thus address two issues at once : eliminate missing terms and not use free-r reflections for this! Pavel
[ccp4bb] AW: [ccp4bb] AW: [ccp4bb] Twinning VS. Disorder
Dear Ian and James, Here I learned something new. I assumed that coherence length would be limited by crystal quality, e.g. mosaicity and microdomains etc. which apparently is not the case. For me, one of the characteristics of twinning is that there is no interference between the twin domains. If the twin domains get so small that there is interference, these phenomenon are usually referred to as lattice translocation disorder, but that might be my personal interpretation. So maybe I focused too narrowly on the word twinning and should have mentioned the lattice translocation disorder. However, in this case the usual twinning options in refinement programs cannot be used, since these assume summation of intensities. Cheers, Herman Von: Ian Tickle [mailto:ianj...@gmail.com] Gesendet: Freitag, 25. April 2014 01:01 An: Schreuder, Herman RD/DE Cc: CCP4BB@JISCMAIL.AC.UK Betreff: Re: [ccp4bb] AW: [ccp4bb] Twinning VS. Disorder Dear Herman On 24 April 2014 22:32, herman.schreu...@sanofi.commailto:herman.schreu...@sanofi.com wrote: The X-ray coherent length is depending on the crystal, not the synchrotron and my gut feeling is that it is at least several hundred unit cells, but here other experts may correct me. I assume you meant that the coherence length is a property of the beam (e.g. for a Cu target source it's related to the lifetime of the excited Cu K-alpha state), not the crystal, e,g, see http://www.aps.anl.gov/Users/Meeting/2010/Presentations/WK2talk_Vartaniants.pdf (slides 8-11). The relevant property of the crystal is the size of the microdomains. You don't get interference because coherence length domain size, i.e. the beam is not coherent over more than 1 domain. This is true for in-house sources synchrotrons, I guess for FELs it's different, i.e. much greater coherence length? This relates to a question I asked on the BB some time ago: if the coherence length is long enough would you start to see the effects of interference in twinned crystals, i.e. would the summation of intensities break down? I defer to the experts on synchrotrons FELs! Cheers -- Ian
Re: [ccp4bb] AW: [ccp4bb] Twinning VS. Disorder
Is the following being neglected? In a crystal with these putative mosaic microdomains, there will be interference between microdomains at their edges/borders (at least), but since most microdomains are probably way smaller than the coherence length of 3-10 microns, presumably all unit cells in domain A interfere with all unit cells in domains B, C, etc, which are in the same coherence volume. In fact, as I said too unclearly in a previous post, as the putative microdomains become smaller and smaller to the limit of one unit cell, they become indistinguishable from unit cell parameter variation. So I am becoming increasingly suspicious about the existence of microdomains, and wonder what hard evidence there is for their existence? As a thought experiment, one can consider the microdomain theory taken to its limit: a powder diffraction image. In powder diffraction, there are so many crystals (read: microdomains) that each spot is manifested at its Bragg angle at every possible radial position on the detector. Mosaicity would be, what, 360 degrees? So, now imagine decreasing the mosaicity to lower values, and one gets progressively shorter arcs which at lower values become spots. Doesn’t this mean that the contribution from microdomain mosaicity should be to make the spots more like arcs, as we sometimes see in terrible diffraction patterns, and not just general broadening of spots? Put another way: mosaicity should broaden spots in the radial direction (arcs), and unit cell parameter variation should produce straight broadening in the direction of the unit cell variation of magnitude proportional to the degree of variation in that direction. JPK From: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] On Behalf Of Ian Tickle Sent: Thursday, April 24, 2014 7:01 PM To: CCP4BB@JISCMAIL.AC.UK Subject: Re: [ccp4bb] AW: [ccp4bb] Twinning VS. Disorder Dear Herman On 24 April 2014 22:32, herman.schreu...@sanofi.commailto:herman.schreu...@sanofi.com wrote: The X-ray coherent length is depending on the crystal, not the synchrotron and my gut feeling is that it is at least several hundred unit cells, but here other experts may correct me. I assume you meant that the coherence length is a property of the beam (e.g. for a Cu target source it's related to the lifetime of the excited Cu K-alpha state), not the crystal, e,g, see http://www.aps.anl.gov/Users/Meeting/2010/Presentations/WK2talk_Vartaniants.pdf (slides 8-11). The relevant property of the crystal is the size of the microdomains. You don't get interference because coherence length domain size, i.e. the beam is not coherent over more than 1 domain. This is true for in-house sources synchrotrons, I guess for FELs it's different, i.e. much greater coherence length? This relates to a question I asked on the BB some time ago: if the coherence length is long enough would you start to see the effects of interference in twinned crystals, i.e. would the summation of intensities break down? I defer to the experts on synchrotrons FELs! Cheers -- Ian
Re: [ccp4bb] AW: [ccp4bb] Twinning VS. Disorder
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Dear Jakob, your Gedankenexperiment on powder diffraction is not correct. You would record a powder diffraction pattern if you rotated a single crystal around the beam axis and record the result on a single image. This rotation does not affect the mosaicity and the mosaicity of a powder sample related only to the mosaicity of the micro crystals present in the powder. You also do not get arcs when reducing the powderness but you start seeing single spots. This can often be observed in the presence of ice rings. Best, Tim On 04/25/2014 09:32 AM, Keller, Jacob wrote: Is the following being neglected? In a crystal with these putative mosaic microdomains, there will be interference between microdomains at their edges/borders (at least), but since most microdomains are probably way smaller than the coherence length of 3-10 microns, presumably all unit cells in domain A interfere with all unit cells in domains B, C, etc, which are in the same coherence volume. In fact, as I said too unclearly in a previous post, as the putative microdomains become smaller and smaller to the limit of one unit cell, they become indistinguishable from unit cell parameter variation. So I am becoming increasingly suspicious about the existence of microdomains, and wonder what hard evidence there is for their existence? As a thought experiment, one can consider the microdomain theory taken to its limit: a powder diffraction image. In powder diffraction, there are so many crystals (read: microdomains) that each spot is manifested at its Bragg angle at every possible radial position on the detector. Mosaicity would be, what, 360 degrees? So, now imagine decreasing the mosaicity to lower values, and one gets progressively shorter arcs which at lower values become spots. Doesn’t this mean that the contribution from microdomain mosaicity should be to make the spots more like arcs, as we sometimes see in terrible diffraction patterns, and not just general broadening of spots? Put another way: mosaicity should broaden spots in the radial direction (arcs), and unit cell parameter variation should produce straight broadening in the direction of the unit cell variation of magnitude proportional to the degree of variation in that direction. JPK From: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] On Behalf Of Ian Tickle Sent: Thursday, April 24, 2014 7:01 PM To: CCP4BB@JISCMAIL.AC.UK Subject: Re: [ccp4bb] AW: [ccp4bb] Twinning VS. Disorder Dear Herman On 24 April 2014 22:32, herman.schreu...@sanofi.commailto:herman.schreu...@sanofi.com wrote: The X-ray coherent length is depending on the crystal, not the synchrotron and my gut feeling is that it is at least several hundred unit cells, but here other experts may correct me. I assume you meant that the coherence length is a property of the beam (e.g. for a Cu target source it's related to the lifetime of the excited Cu K-alpha state), not the crystal, e,g, see http://www.aps.anl.gov/Users/Meeting/2010/Presentations/WK2talk_Vartaniants.pdf (slides 8-11). The relevant property of the crystal is the size of the microdomains. You don't get interference because coherence length domain size, i.e. the beam is not coherent over more than 1 domain. This is true for in-house sources synchrotrons, I guess for FELs it's different, i.e. much greater coherence length? This relates to a question I asked on the BB some time ago: if the coherence length is long enough would you start to see the effects of interference in twinned crystals, i.e. would the summation of intensities break down? I defer to the experts on synchrotrons FELs! Cheers -- Ian - -- - -- Dr Tim Gruene Institut fuer anorganische Chemie Tammannstr. 4 D-37077 Goettingen GPG Key ID = A46BEE1A -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.12 (GNU/Linux) Comment: Using GnuPG with Icedove - http://www.enigmail.net/ iD8DBQFTWiJpUxlJ7aRr7hoRAvhpAKCWt3PwAQsPnUgMlHjYoGS/7lVlGACglWpz K+rZPikLZBwe+CrK29WhBnc= =4a9F -END PGP SIGNATURE-
Re: [ccp4bb] coot problems to decrease R FREE
Hi Pavel, I agree that you bias R-free after the real-space refinement well, ok, isn't it enough to realize that this is bad and should be avoided ? (I guess we all know we should never bias Rfree!) My point was that we normally do not calculate R-free after real- space refinement, It's not about whether you compute something or not. It's about whether you expose free-r reflections to refinement and to current model, given that they should never ever see any refinement or model or optimization under any circumstances. Otherwise they immediately become non-free. Yet we do optimise weights against R-free, both in phenix.refine (from the manual: phenix.refine uses automatic procedure to determine the weights between X-ray target and stereochemistry or ADP restraints. To optimize these weights (that is to find those resulting in lowest Rfree factors)) and in PDB_REDO (well against LL-free, but that is the same test set). And we do not go the distance to use something like an R-sleep set to keep a set of reflections really independent. Here we are introducing bias as well, but we seem to have accepted that this is not a big problem. Or we chose to ignore it. but after reciprocal space refinement. Here the bias is removed again. Hm.. How you know this? I guess it requires a great deal of effort to prove this!. We use this whenever we are forced to use a new test set. E.g. when doing molecular replacement into an isomorphous cell without access to the original test set (this has been discussed on the BB a few times) or when doing k-fold cross validation. Whenever a new test set in chosen, enough model tuning (i.e. refinement) is needed to make the model independent of the test set (you may need to perturb the model somewhat by resetting the B-factors or by giving the coordinates a nudge). At refinement convergence the maximum amount of tuning is attained. And at true convergence, the starting point doesn't matter (although local minima complicate matters). With k-fold cross validation we can actually show that there is no obvious bias: Say we build and refine a model with test set 0 left out, then any further refinement with k different test sets left out should start out with R-free substantially biased for any set i!=0, that is R-free(i) R-free(0). This is indeed what we see. Now if the refinements converge and there still is bias then we would still get R-free(i) R-free(0) for all sets i!=0. This is not what we see (at least not for most PDB_REDO test cases). Of course, these are examples of the worst sort of test set bias. The bias introduced in real-space refinement is more subtle. At least that is what I assume, but this is something you said you could quantify. A proper estimate of the problem (rather than a principle we don't quite adhere to all the time anyway), would really help. Cheers, Robbie In practical terms we have to choose the best option: 1) Refine against maps with missing reflections and the possibility of artefacts that lead to suboptimal results. This keeps R-free unbiased, even directly after real-space refinement. 2) Refine against maps with test reflections set to Fcalc. This biases against the current model, but should have less artefacts. This too keeps R- free unbiased directly after real-space refinement. 3) Refine against maps with all reflections. This should give the best fitting results, but does introduce bias to the test set. However, this bias These are all valid points. However they do not advocate for biasing Rfree. Yuo can always fill in missing reflections with something else (different from your genuine free-r set of reelections) and thus address two issues at once : eliminate missing terms and not use free-r reflections for this! Pavel
[ccp4bb] Postdoctoral Position Available at Rutgers University
A postdoctoral position is available in the Department of Microbiology and Molecular Genetics at Rutgers University - New Jersey Medical School to study structural mechanisms regulating bacterial signaling and perform SAR studies of antimicrobial compounds. In addition to a recent Ph.D., the ideal candidate will have a strong background in macromolecular X-ray crystallography, biochemistry, protein purification, and molecular biology. A working knowledge of bacterial genetics and computational modeling of protein-ligand interactions is a plus. Candidates with demonstrated success in another field and a strong commitment to learning the crystallographic process will also be considered. Please submit: CV and at least 2 letters of reference by email to matthew.neidi...@rutgers.edu with the subject line “Postdoctoral Position Applicant” Matthew B. Neiditch, Ph.D. Associate Professor Dept. of Microbiology and Molecular Genetics New Jersey Medical School Rutgers, The State University of New Jersey 225 Warren St., Room E450U Newark, NJ 07103
Re: [ccp4bb] AW: [ccp4bb] Twinning VS. Disorder
your Gedankenexperiment on powder diffraction is not correct. You would record a powder diffraction pattern if you rotated a single crystal around the beam axis and record the result on a single image. If you wanted to do it with a single crystal, you would have to rotate the crystal through all possible rotations in 3d, not just around the axis of the beam, because you would then miss all the reflections which were not in the diffraction condition at that phi angle. I agree that it could be done this way (not sure why this is important though.) This rotation does not affect the mosaicity and the mosaicity of a powder sample related only to the mosaicity of the micro crystals present in the powder. You also do not get arcs when reducing the powderness but you start seeing single spots. This can often be observed in the presence of ice rings. You are talking about powderness, which I would guess is a measure of the completeness of the sampling of all possible orientations of the constituent crystals, and I agree with what you say would happen. I said, however, mosaicity, which is a measure of the breadth of the distribution of the orientation angles of the microdomains/microcrystals. By decreasing powderness, one would do nothing to mosaicity. If one could arrange the microdomains into some range of orientation angles, one would reduce the mosaicity, and get arcs. I wish I had a picture of an arc-containing diffraction pattern--I've seen them from time to time, and they're always of course bad news. Anyone on the list have such a diffraction pattern handy? JPK On 04/25/2014 09:32 AM, Keller, Jacob wrote: Is the following being neglected? In a crystal with these putative mosaic microdomains, there will be interference between microdomains at their edges/borders (at least), but since most microdomains are probably way smaller than the coherence length of 3-10 microns, presumably all unit cells in domain A interfere with all unit cells in domains B, C, etc, which are in the same coherence volume. In fact, as I said too unclearly in a previous post, as the putative microdomains become smaller and smaller to the limit of one unit cell, they become indistinguishable from unit cell parameter variation. So I am becoming increasingly suspicious about the existence of microdomains, and wonder what hard evidence there is for their existence? As a thought experiment, one can consider the microdomain theory taken to its limit: a powder diffraction image. In powder diffraction, there are so many crystals (read: microdomains) that each spot is manifested at its Bragg angle at every possible radial position on the detector. Mosaicity would be, what, 360 degrees? So, now imagine decreasing the mosaicity to lower values, and one gets progressively shorter arcs which at lower values become spots. Doesn’t this mean that the contribution from microdomain mosaicity should be to make the spots more like arcs, as we sometimes see in terrible diffraction patterns, and not just general broadening of spots? Put another way: mosaicity should broaden spots in the radial direction (arcs), and unit cell parameter variation should produce straight broadening in the direction of the unit cell variation of magnitude proportional to the degree of variation in that direction. JPK From: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] On Behalf Of Ian Tickle Sent: Thursday, April 24, 2014 7:01 PM To: CCP4BB@JISCMAIL.AC.UK Subject: Re: [ccp4bb] AW: [ccp4bb] Twinning VS. Disorder Dear Herman On 24 April 2014 22:32, herman.schreu...@sanofi.commailto:herman.schreu...@sanofi.com wrote: The X-ray coherent length is depending on the crystal, not the synchrotron and my gut feeling is that it is at least several hundred unit cells, but here other experts may correct me. I assume you meant that the coherence length is a property of the beam (e.g. for a Cu target source it's related to the lifetime of the excited Cu K-alpha state), not the crystal, e,g, see http://www.aps.anl.gov/Users/Meeting/2010/Presentations/WK2talk_Vartan iants.pdf (slides 8-11). The relevant property of the crystal is the size of the microdomains. You don't get interference because coherence length domain size, i.e. the beam is not coherent over more than 1 domain. This is true for in-house sources synchrotrons, I guess for FELs it's different, i.e. much greater coherence length? This relates to a question I asked on the BB some time ago: if the coherence length is long enough would you start to see the effects of interference in twinned crystals, i.e. would the summation of intensities break down? I defer to the experts on synchrotrons FELs! Cheers -- Ian - -- - -- Dr Tim Gruene Institut fuer anorganische Chemie Tammannstr. 4 D-37077 Goettingen GPG Key ID = A46BEE1A -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.12 (GNU/Linux)
Re: [ccp4bb] PyMol and Schrodinger
Thank you Piotr, this is very useful to know. Cheers, Eugene On 24 Apr 2014, at 22:18, Piotr Sliz wrote: Eugene, SBGrid Consortium supports certain paid programs such as Pymol, Schrodinger, and Geneious, and academic laboratories have access to a limited pool of tokens. For more serious computations it might be easier to get an individual license, yet for starters the SBGrid shared-token library is certainly a viable option. Access to the commercial tokens is included in SBGrid membership, though limited depending on your geographic location and affiliation. Please email h...@sbgrid.org for a more detailed description of this program. Kind Regards, Piotr -- Scanned by iCritical.
[ccp4bb] AW: [ccp4bb] Twinning VS. Disorder
Dear James, I have a slightly different way to think about transverse coherence. I heard Mark Sutton give a talk about this at the APS a few years ago, and here are graphics from a similar talk given by Alec Sandy at BNL: http://www.bnl.gov/nsls2/workshops/docs/XPCS/XPCS_Sandy.ppt The eqn. he gives on figure 8 is this: L(coherence) = Lambda*(source-to-observation-point-distance) / 2*pi*sigma(source) This is the wavelength divided by the angle subtended by the source viewed by the observer, with a 2.pi in there for some reason. The way I explain it (though I cannot derive it this way) is that as you view the source, consider that your eye sees a ray coming from the top of the source, and one coming from the bottom. As you move your eye up and down, the two rays will slide back and forth against one another. The coherence length is how far you move your eye to have them slip about 1/4 of a wave. In Mark's APS example you see the horizontal is 7 microns, the vertical is 200. This reflects the fact that the typical synchrotron source is much wider than it is high. Bob From: CCP4 bulletin board [CCP4BB@JISCMAIL.AC.UK] on behalf of James Holton [jmhol...@lbl.gov] Sent: Thursday, April 24, 2014 6:59 PM To: CCP4BB@JISCMAIL.AC.UK Subject: Re: [ccp4bb] AW: [ccp4bb] Twinning VS. Disorder There are two kinds of coherence length: transverse and longitudinal. Longitudinal coherence is often quoted as delta-lambda/lambda, which is easy to calculate but unfortunately completely irrelevant for diffraction from crystals. If it weren't then Laue diffraction wouldn't produce spots. Transverse coherence tends to be around 3-10 microns with 1 A x-rays, depending on the detector distance. Yes, that's right, the detector distance. Longer detector distances give you a bigger coherence length, especially when the source is very far away, like it is at a synchrotron. How this happens is easiest to picture if you consider the simplest possible diffraction situation: a point source of x-rays, two atoms, and a detector. As long as the atoms are very close together relative to the distances from the sample to the source and the detector, then you have the far field diffraction situation. This is where both atoms are within the coherence length, Bragg's diagram for Bragg's Law holds: parallel incoming rays and parallel outgoing rays. But what if the atoms are very far apart? Obviously, the scattering from two atoms on different sides of the room will just add as intensities. And if they are very close together, then Bragg's Law holds and they scatter coherently. What most people think of as the coherence length is the point of transition between these two kinds of scattering. This point is rather conveniently defined as the distance between two atoms when the path from the source to one atom to a given detector pixel becomes 0.5 wavelengths longer than the same path through the other atom. As long as both atoms lie in the Bragg plane (that's the plane perpendicular to the s vector, which is the vector difference between the incoming and outgoing beam directions), the far-field approximation tells us they should also be in phase, but if they are far enough apart the 0.5 A change in total path length is enough to change the scattering completely from constructive to destructive interference. In ordinary optics, this is called the edge of the first Fresnel zone. So, if your source is very far away, emitting 1 A x-rays, and your detector is 1 meter away, then moving one atom 10 microns away from the centerline of the beam makes the path from that atom to the detector 1-sqrt(1^2+10e-6^2) = 0.5 A longer. So that implies the coherence length is 10 microns. But if the detector is only 100 mm away, that gives you 0.1-sqrt(0.1^2+3e-6^2) = 0.5 A, so 3 um is the coherence length. Of course, this is for the ideal case of a point source very far away. In reality finite beam divergence will mess up the coherence inasmuch as a divergent source looks like an array of sources all viewing the sample through a pinhole. What you then get on the detector is the sum of the patterns from all those sources, so the coherence is not as clean. That is, you don't see the Fourier transform of the crystal shape in every spot. Mosaic spread also messes up coherence in this way. Some might even define the mosaic domain size as the inverse of the effective coherence length. But, the long and short of all this is that as long as your detector pixels are bigger than the coherence length the coherence doesn't really matter. Hope that makes sense, -James Holton MAD Scientist On Thu, Apr 24, 2014 at 2:32 PM, herman.schreu...@sanofi.commailto:herman.schreu...@sanofi.com wrote: Dear Chen, Twinning can be thought of as of two or more macro-crystals glued or grown together. The reason that the reflections often overlap is that they
Re: [ccp4bb] coot problems to decrease R FREE
Hi Robbie, I agree that you bias R-free after the real-space refinement well, ok, isn't it enough to realize that this is bad and should be avoided ? (I guess we all know we should never bias Rfree!) My point was that we normally do not calculate R-free after real- space refinement, It's not about whether you compute something or not. It's about whether you expose free-r reflections to refinement and to current model, given that they should never ever see any refinement or model or optimization under any circumstances. Otherwise they immediately become non-free. Yet we do optimise weights against R-free, both in phenix.refine (from the manual: phenix.refine uses automatic procedure to determine the weights between X-ray target and stereochemistry or ADP restraints. To optimize these weights (that is to find those resulting in lowest Rfree factors)) and in PDB_REDO (well against LL-free, but that is the same test set). And we do not go the distance to use something like an R-sleep set to keep a set of reflections really independent. Here we are introducing bias as well, but we seem to have accepted that this is not a big problem. Or we chose to ignore it. This is all true. There is a subtle line, however, between actively refining your model against free reflections (and therefore against Rfree) in real or reciprocal spaces, and using Rfree at discrete points of time to make certain decisions, like visually inspect the numbers and say I like this model more because it has lower Rfree, or choose weight because one yields a lower Rfree than the other. Technically, I agree, it is also using test reflections, and it would be the best if sleep set of reflections is used for this instead! From practical side, it's easy to demonstrate that if you start with a model and data set composed of work and free reflections, and having RfreeRwork with some gap, then compute two maps: one using only work reflections, and another using all work+test. Then refine model against the two maps. In case of refinement against the first map (calculated using work reflections only), both Rfree and Rwork will change, RfreeRwork will remain true and the gap will be maintained. In case of refinement against the second map (calculated using all, work+test, reflections), both Rfree and Rwork will change, and Rfree will *very* quickly approach Rwork, so Rwork~Rfree. I've seen it numerous times when developing real-space refinement in Phenix, and I think I can still find the script that demonstrates specifically this scenario. All the best, Pavel
[ccp4bb] what happens when freezer goes down
Dear colleagues, I know this is not related to ccp4 but I am in need of an answer and many of you work with cells etc. My building had a major malfunction of electricity and the power backup did not kick in. My -80C freezer was without power for over 24 hours until I found out. Because it is small, it goes fast to room temp. I had many glycerol stocks in it with cells, cells with plasmids etc. as well as cell pellets. I am trying to rescue things. My question is what happens to cell pellets. I had many as I like to start purification at the cell pellet level. Are these destroyed when they go to room temp for 24 hours or are they ok? Thanks. Jackie Vitali Cleveland State University
[ccp4bb] CCP4-6.4.0 Update 013
Dear CCP4 Users An update for the CCP4-6.4.0 series has just been released, consisting of the following changes * qtpisa and pisa - New features + concentration dependence of predicted oligomeric states to aid the identification of biological units in protein's working conditions + multi-parameter interaction radar for more consistent discrimination between biologically-relevant interfaces and crystal contacts + help support (qtpisa only) Note that auto-updates work only with CCP4 6.4.0 series, therefore please upgrade if necessary. The Update Manager is now included in the package so you do not need to install it separately. In addition, all available updates will be installed automatically if you are using Setup Manager for CCP4 installation. Please report any bugs to c...@stfc.ac.uk. Many thanks for using CCP4. The CCP4 Core Team-- Scanned by iCritical.
Re: [ccp4bb] what happens when freezer goes down
Jackie, We grow cells routinely and freeze pellets after fermentation. In general, proteins are fairly stable until you break cells so you are probably ok unless your protein is very heat labile and it sat at room temperature for hours. However as I mentioned, there is a kind of buffering from the cell that can stabilize the protein until you are ready to use it. The glycerol stocks are probably ok as well since you need just a small inoculum to get your culture growing. The plasmids may be the most affected, but you can sequence if you have aberrant expression. Best, David -- David L. Blum, Ph.D. Director, Bioexpression and Fermentation Facility Department of Biochemistry and Molecular Biology University of Georgia 120 Green Street room A414A Athens, GA 30602 http://bff.uga.edu/ b...@uga.edu On Fri, Apr 25, 2014 at 12:43 PM, Jacqueline Vitali jackie.vit...@gmail.com wrote: Dear colleagues, I know this is not related to ccp4 but I am in need of an answer and many of you work with cells etc. My building had a major malfunction of electricity and the power backup did not kick in. My -80C freezer was without power for over 24 hours until I found out. Because it is small, it goes fast to room temp. I had many glycerol stocks in it with cells, cells with plasmids etc. as well as cell pellets. I am trying to rescue things. My question is what happens to cell pellets. I had many as I like to start purification at the cell pellet level. Are these destroyed when they go to room temp for 24 hours or are they ok? Thanks. Jackie Vitali Cleveland State University
Re: [ccp4bb] AW: [ccp4bb] Twinning VS. Disorder
Thanks to all—I’ve got the paper now JPK From: Keller, Jacob Sent: Friday, April 25, 2014 1:58 PM To: 'Oliver Zeldin' Cc: CCP4BB@jiscmail.ac.uk Subject: RE: [ccp4bb] AW: [ccp4bb] Twinning VS. Disorder Does anyone know of a place where one can obtain this reference for free? I would contact Darwin himself, but I suspect he wouldn’t write back. I think this is the original paper proposing the mosaic block model, and I’d really like to see his reasoning. Darwin, C. G. (1922). Philos. Mag. 43, 800±829. The reflexion of X-rays from imperfect crystals JPK From: oliver.zel...@gmail.commailto:oliver.zel...@gmail.com [mailto:oliver.zel...@gmail.com] On Behalf Of Oliver Zeldin Sent: Friday, April 25, 2014 1:03 PM To: Keller, Jacob Cc: CCP4BB@jiscmail.ac.ukmailto:CCP4BB@jiscmail.ac.uk Subject: Re: [ccp4bb] AW: [ccp4bb] Twinning VS. Disorder Dear Jacob, In terms of the effect of crystal (lattice) defects on diffraction spot profiles, there are two great papers by Colin Nave that discuss this: http://journals.iucr.org/d/issues/1998/05/00/issconts.html and http://journals.iucr.org/d/issues/1998/05/00/issconts.html . There is also this paper on the nature of mosaic micro-domains: http://journals.iucr.org/d/issues/2000/08/00/en0024/en0024.pdf. I am sure there must be other references for the 'nature' of lattice disorder, and it anyone can point to them, I would be grateful. Cheers, Oliver On Fri, Apr 25, 2014 at 6:20 AM, Keller, Jacob kell...@janelia.hhmi.orgmailto:kell...@janelia.hhmi.org wrote: your Gedankenexperiment on powder diffraction is not correct. You would record a powder diffraction pattern if you rotated a single crystal around the beam axis and record the result on a single image. If you wanted to do it with a single crystal, you would have to rotate the crystal through all possible rotations in 3d, not just around the axis of the beam, because you would then miss all the reflections which were not in the diffraction condition at that phi angle. I agree that it could be done this way (not sure why this is important though.) This rotation does not affect the mosaicity and the mosaicity of a powder sample related only to the mosaicity of the micro crystals present in the powder. You also do not get arcs when reducing the powderness but you start seeing single spots. This can often be observed in the presence of ice rings. You are talking about powderness, which I would guess is a measure of the completeness of the sampling of all possible orientations of the constituent crystals, and I agree with what you say would happen. I said, however, mosaicity, which is a measure of the breadth of the distribution of the orientation angles of the microdomains/microcrystals. By decreasing powderness, one would do nothing to mosaicity. If one could arrange the microdomains into some range of orientation angles, one would reduce the mosaicity, and get arcs. I wish I had a picture of an arc-containing diffraction pattern--I've seen them from time to time, and they're always of course bad news. Anyone on the list have such a diffraction pattern handy? JPK On 04/25/2014 09:32 AM, Keller, Jacob wrote: Is the following being neglected? In a crystal with these putative mosaic microdomains, there will be interference between microdomains at their edges/borders (at least), but since most microdomains are probably way smaller than the coherence length of 3-10 microns, presumably all unit cells in domain A interfere with all unit cells in domains B, C, etc, which are in the same coherence volume. In fact, as I said too unclearly in a previous post, as the putative microdomains become smaller and smaller to the limit of one unit cell, they become indistinguishable from unit cell parameter variation. So I am becoming increasingly suspicious about the existence of microdomains, and wonder what hard evidence there is for their existence? As a thought experiment, one can consider the microdomain theory taken to its limit: a powder diffraction image. In powder diffraction, there are so many crystals (read: microdomains) that each spot is manifested at its Bragg angle at every possible radial position on the detector. Mosaicity would be, what, 360 degrees? So, now imagine decreasing the mosaicity to lower values, and one gets progressively shorter arcs which at lower values become spots. Doesn’t this mean that the contribution from microdomain mosaicity should be to make the spots more like arcs, as we sometimes see in terrible diffraction patterns, and not just general broadening of spots? Put another way: mosaicity should broaden spots in the radial direction (arcs), and unit cell parameter variation should produce straight broadening in the direction of the unit cell variation of magnitude proportional to the degree of variation in that direction. JPK From: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UKmailto:CCP4BB@JISCMAIL.AC.UK] On Behalf Of Ian
Re: [ccp4bb] NEW! Try depositing X-ray data using the wwPDB deposition system
I was getting very excited here for a second. Deposition of X-ray data? Is this finally possible? Raw diffraction data, in other words? Image files? Sadly, the text of the email doesn't support that conclusion. Andreas On 25/04/2014 8:16, Rachel Kramer Green wrote: Depositing a new X-ray crystal structure to the PDB archive?Try the new and improved wwPDB Deposition system at http://deposit.wwpdb.org/deposition/. Since the new system went live on January 27^th , 2014, ~700 structures have been deposited and promptly annotated.Detailed information on the system can be found at http://www.wwpdb.org/system_info.html.The new system was developed to allow the wwPDB partners to meet the evolving needs of the scientific community over the next decade, including support for very large systems, complex chemistry, and joint use of multiple experimental methods. The system replaces all current deposition and annotation systems in use at the wwPDB deposition centers, and will lead to improved efficiency and consistency. New or enhanced features of the deposition system include the generation of X-ray validation reports (following the recommendations of the wwPDB X-ray Validation Task Force), improved capture and review of ligand information, the ability to replace coordinate and/or experimental data files pre- and post-submission, improved communication process between depositors and wwPDB curators, and the ability to preview and download the PDBx/mmCIF entry file prior to submission. Depositors will have the option to use the new system or one of the legacy deposition tools (ADIT, AutoDep) for most of 2014. After the transition to the new system, the legacy tools will be available for a limited period of time to complete any unfinished deposition sessions. The wwPDB continues to improve the deposition interface and supporting functionality as it receives more submissions and input.Questions, comments, and suggestions should be sent to: http://deposit-feedback.wwpdb.org_._ -- Rachel Kramer Green, Ph.D. RCSB PDB kra...@rcsb.rutgers.edu *New!*Deposit X-ray data with the wwPDB at: http://deposit.wwpdb.org/deposition (NMR and 3DEM coming soon). ___ Twitter: https://twitter.com/#!/buildmodels Facebook: http://www.facebook.com/RCSBPDB -- Andreas Förster Crystallization and X-ray Facility Manager Centre for Structural Biology Imperial College London
[ccp4bb] anomalous signal
Dear all written below is the log file of an anomalous data processed through SHELXC..my question is ..what is the strength of anomalous signal ?? as it is said For zero signal d'/sig and d/sig should be about 0.80. Then in the present case is there really a signal or can be assumed no signal..we are expecting one Ca atom bound to the protein at its active site..the redundancy of the data is 11.6..with this signal strength can we assume Ca to be present there or whatever little anomalous if present is due to something elseor there is no signal at all ??... Resl. Inf - 8.0 - 6.0 - 5.0 - 4.0 - 3.8 - 3.6 - 3.4 - 3.2 - 3.0 - 2.8 - 2.60 N(data) 375 493 580 1319 450 538 679 866 1081 1414 1709 I/sig58.8 38.6 32.6 38.3 27.7 27.2 21.9 18.4 12.6 9.5 6.1 %Complete 94.7 99.0 99.3 99.5 100.0 99.6 99.7 99.8 99.6 99.6 90.9 d/sig 1.65 1.27 1.18 1.25 1.19 1.12 1.11 1.11 0.97 1.02 1.05 -- Regards Faisal School of Life Sciences JNU
[ccp4bb] anomalous signal
Dear all sorry about my previous mail where i forgot to mention that the data was collected on home source at Cuk alpha and at 1.54A. written below is the log file of an anomalous data processed through SHELXC..my question is ..what is the strength of anomalous signal ?? as it is said For zero signal d'/sig and d/sig should be about 0.80. Then in the present case is there really a signal or can be assumed no signal..we are expecting one Ca atom bound to the protein at its active site..the redundancy of the data is 11.6..with this signal strength can we assume Ca to be present there or whatever little anomalous if present is due to something elseor there is no signal at all ??... Resl. Inf - 8.0 - 6.0 - 5.0 - 4.0 - 3.8 - 3.6 - 3.4 - 3.2 - 3.0 - 2.8 - 2.60 N(data) 375 493 580 1319 450 538 679 866 1081 1414 1709 I/sig58.8 38.6 32.6 38.3 27.7 27.2 21.9 18.4 12.6 9.5 6.1 %Complete 94.7 99.0 99.3 99.5 100.0 99.6 99.7 99.8 99.6 99.6 90.9 d/sig 1.65 1.27 1.18 1.25 1.19 1.12 1.11 1.11 0.97 1.02 1.05 -- Regards Faisal School of Life Sciences JNU
Re: [ccp4bb] anomalous signal
d/sig should be above 0.80 There seems to be plenty of signal there with all values above 1.02. We have solved structures with less multiplicity and lower d/sig. There is a different criteria of signal for when you know the positions of the anomalous substructure atoms and when you need to find the positions of the anomalous substructure atoms. As for no signal, I think I am on record that there is always an anomalous signal. :) But can you detect it? Jim From: CCP4 bulletin board [CCP4BB@JISCMAIL.AC.UK] on behalf of Faisal Tarique [faisaltari...@gmail.com] Sent: Friday, April 25, 2014 4:06 PM To: CCP4BB@JISCMAIL.AC.UK Subject: [ccp4bb] anomalous signal Dear all sorry about my previous mail where i forgot to mention that the data was collected on home source at Cuk alpha and at 1.54A. written below is the log file of an anomalous data processed through SHELXC..my question is ..what is the strength of anomalous signal ?? as it is said For zero signal d'/sig and d/sig should be about 0.80. Then in the present case is there really a signal or can be assumed no signal..we are expecting one Ca atom bound to the protein at its active site..the redundancy of the data is 11.6..with this signal strength can we assume Ca to be present there or whatever little anomalous if present is due to something elseor there is no signal at all ??... Resl. Inf - 8.0 - 6.0 - 5.0 - 4.0 - 3.8 - 3.6 - 3.4 - 3.2 - 3.0 - 2.8 - 2.60 N(data) 375 493 580 1319 450 538 679 866 1081 1414 1709 I/sig58.8 38.6 32.6 38.3 27.7 27.2 21.9 18.4 12.6 9.5 6.1 %Complete 94.7 99.0 99.3 99.5 100.0 99.6 99.7 99.8 99.6 99.6 90.9 d/sig 1.65 1.27 1.18 1.25 1.19 1.12 1.11 1.11 0.97 1.02 1.05 -- Regards Faisal School of Life Sciences JNU
Re: [ccp4bb] Systematically shifted peak in 2Fo-Fc and BDF maps
Thanks to all who replied to my message. The most logical explanation to the observed phenomenon is likely the following. During crystallization, Sr ions from the crystallization solution kick out a fraction of the Mg(OH)6 ions from the binding site in question. The cavity of the binding site is big and the Sr ion binds at a slightly different position compared to the Mg(OH)6 ion. The BDF map shows the shifted position of the Sr ion, whereas the 2Fo-Fc map shows the position of the Mg(OH)6 ion. This is possible because the occupancy of the Sr ion at that site is low - the height of the Sr peak in the BDF map is less than 15% of the maximal peak height - and the contribution of the Sr ion to the 2Fo-Fc map is negligibly small. The “normal” (2Fo-Fc) scattering of the Sr ion is essentially completely masked by the Mg(OH)6 ion scattering because the two overlap (they are 0.47 A apart) and the latter has a 5-6 times higher occupancy. Eleanor’s and Phoebe Rice’s comments were extremely useful in solving this puzzle (at least for me - we will see if the referees agree that this is a solution). This particular comment by Phoebe was really insightful considering she did not see the actual data but only read my somewhat incomplete description: “Could it be a mix of species - partly Sr and partly some non-anomalous-scattering entity that sits in a slightly shifted position?” CCP4BB is an incredible resource full of really clever people! Thanks to all again! Petr On Apr 24, 2014, at 11:39, Petr Leiman petr.lei...@epfl.chmailto:petr.lei...@epfl.ch wrote: Dear All, We are looking for an explanation for a very strange observation. Problem: We have two fully independent data sets (two different crystals), in which the Bijvoet Difference Fourier map peak of one particular metal site is shifted by 0.47 A from its position in the 2Fo-Fc map. Relevant information: The resolution of both data sets is 1.5-1.6 A. The 2Fo-Fc and BDP maps are calculated using the same phases. The metal ion is water hydrated and all the details are crystal clear in both 2Fo-Fc maps. The crystals are grown in the presence of Sr and the data sets are collected at the Sr K-edge. There are many other Sr sites and all strong peaks in the BDF map overlap with 2Fo-Fc map peaks perfectly. The Sr site in question is not the strongest, but it is well above the noise level of the BDF map. Additional information: The weird site is actually fully buried inside an internal cavity (it is surrounded by protein atoms from all sides), but a Sr atom is able to diffuse into this cavity somehow. All other Sr sites are on the surface of the protein. Any thoughts about why a non-noise BDF map peak would not overlap with a 2Fo-Fc map peak are welcome! Thank you very much, Petr -- Petr Leiman EPFL BSP 415 CH-1015 Lausanne Switzerland Office: +41 21 69 30 441 Mobile: +41 79 538 7647 Fax: +41 21 69 30 422
Re: [ccp4bb] AW: [ccp4bb] Twinning VS. Disorder
Thank you all for your comments/references and now I have a better understanding of what could be actually happening. But I have a feeling that disorder, where the twin domains can interfere with each other, is not actually so unusual. And in some cases MR might be possible to reveal a partial structure. (I ran into papers like this before but I couldn't think of one at the top of my head.) Why there is not a package out there to allow refinement against such data? I feel that if you have phase information it is possible to model the translation between different micro-domains... Sincerely, Chen On Apr 25, 2014, at 3:21 PM, Keller, Jacob kell...@janelia.hhmi.org wrote: Thanks to all—I’ve got the paper now JPK From: Keller, Jacob Sent: Friday, April 25, 2014 1:58 PM To: 'Oliver Zeldin' Cc: CCP4BB@jiscmail.ac.uk Subject: RE: [ccp4bb] AW: [ccp4bb] Twinning VS. Disorder Does anyone know of a place where one can obtain this reference for free? I would contact Darwin himself, but I suspect he wouldn’t write back. I think this is the original paper proposing the mosaic block model, and I’d really like to see his reasoning. Darwin, C. G. (1922). Philos. Mag. 43, 800±829. The reflexion of X-rays from imperfect crystals JPK From: oliver.zel...@gmail.com [mailto:oliver.zel...@gmail.com] On Behalf Of Oliver Zeldin Sent: Friday, April 25, 2014 1:03 PM To: Keller, Jacob Cc: CCP4BB@jiscmail.ac.uk Subject: Re: [ccp4bb] AW: [ccp4bb] Twinning VS. Disorder Dear Jacob, In terms of the effect of crystal (lattice) defects on diffraction spot profiles, there are two great papers by Colin Nave that discuss this: http://journals.iucr.org/d/issues/1998/05/00/issconts.html and http://journals.iucr.org/d/issues/1998/05/00/issconts.html . There is also this paper on the nature of mosaic micro-domains: http://journals.iucr.org/d/issues/2000/08/00/en0024/en0024.pdf. I am sure there must be other references for the 'nature' of lattice disorder, and it anyone can point to them, I would be grateful. Cheers, Oliver On Fri, Apr 25, 2014 at 6:20 AM, Keller, Jacob kell...@janelia.hhmi.org wrote: your Gedankenexperiment on powder diffraction is not correct. You would record a powder diffraction pattern if you rotated a single crystal around the beam axis and record the result on a single image. If you wanted to do it with a single crystal, you would have to rotate the crystal through all possible rotations in 3d, not just around the axis of the beam, because you would then miss all the reflections which were not in the diffraction condition at that phi angle. I agree that it could be done this way (not sure why this is important though.) This rotation does not affect the mosaicity and the mosaicity of a powder sample related only to the mosaicity of the micro crystals present in the powder. You also do not get arcs when reducing the powderness but you start seeing single spots. This can often be observed in the presence of ice rings. You are talking about powderness, which I would guess is a measure of the completeness of the sampling of all possible orientations of the constituent crystals, and I agree with what you say would happen. I said, however, mosaicity, which is a measure of the breadth of the distribution of the orientation angles of the microdomains/microcrystals. By decreasing powderness, one would do nothing to mosaicity. If one could arrange the microdomains into some range of orientation angles, one would reduce the mosaicity, and get arcs. I wish I had a picture of an arc-containing diffraction pattern--I've seen them from time to time, and they're always of course bad news. Anyone on the list have such a diffraction pattern handy? JPK On 04/25/2014 09:32 AM, Keller, Jacob wrote: Is the following being neglected? In a crystal with these putative mosaic microdomains, there will be interference between microdomains at their edges/borders (at least), but since most microdomains are probably way smaller than the coherence length of 3-10 microns, presumably all unit cells in domain A interfere with all unit cells in domains B, C, etc, which are in the same coherence volume. In fact, as I said too unclearly in a previous post, as the putative microdomains become smaller and smaller to the limit of one unit cell, they become indistinguishable from unit cell parameter variation. So I am becoming increasingly suspicious about the existence of microdomains, and wonder what hard evidence there is for their existence? As a thought experiment, one can consider the microdomain theory taken to its limit: a powder diffraction image. In powder diffraction, there are so many crystals (read: microdomains) that each spot is manifested at its Bragg angle at every possible radial position on the detector. Mosaicity would be, what, 360 degrees? So, now imagine
Re: [ccp4bb] what happens when freezer goes down
Hi, My condolences. This is not a fun experience! Your plasmids will be OK :) Glycerol stocks may be rescued as long as after going to room temperature they did NOT get frozen again. Re-freeze will greatly reduce viability - but even after one re-freeze you should be able to streak the glycerol stocks onto appropriate plates and get a few (or more) colonies. Notably this does not mean that expression from these rescued glycerols will be good; furthermore there are quite a few protein expression constructs in our labs that do not store well as glycerol stocks - fresh DNA transformation is the only way to get consistent expression for them. Proteins in frozen cell pellets: that's iffy. The outcome greatly depends on the protein itself - for instance I had several proteins that had activity loss issues even when stored in LN2. In a -80 freezer they would lose 40-50% activity in a week, and a single freeze/thaw/freeze would kill them instantly (even in cell pellets). Then again, other proteins are fine 'no matter what'. In cases like yours it really helps to have some sort of parameter you can follow (activity, or something). Frozen protein samples - can be even more iffy than pellets. Pellets are generally cheap to make, though. Unless they are isotope-labeled or suchlike. Artem - Cosmic Cats approve of this message On Fri, Apr 25, 2014 at 11:43 AM, Jacqueline Vitali jackie.vit...@gmail.com wrote: Dear colleagues, I know this is not related to ccp4 but I am in need of an answer and many of you work with cells etc. My building had a major malfunction of electricity and the power backup did not kick in. My -80C freezer was without power for over 24 hours until I found out. Because it is small, it goes fast to room temp. I had many glycerol stocks in it with cells, cells with plasmids etc. as well as cell pellets. I am trying to rescue things. My question is what happens to cell pellets. I had many as I like to start purification at the cell pellet level. Are these destroyed when they go to room temp for 24 hours or are they ok? Thanks. Jackie Vitali Cleveland State University
