Dear Andrey, Due to our company policy, I cannot view the dropbox image, but the packing you describe is exactly what Edward Barry described and what I expected. Two molecules obey the 2-fold symmetry and one molecule is lying on a 2-fold with two possible superimposed orientations. The question is: are there frequent switches between both conformations such that X-rays diffracted from both conformations are interfering, or are there just larger chunks of crystal with either conformation A or B, behaving like separate crystals? In the first case, one has a crystal packing disorder and, as Eleanor was hinting at, some of the diffraction spots may look funny or smeared, in the second case one has twinning.
I do not know the exact method sfcheck uses, but normally to detect twinning, programs do not look at symmetry but look at intensity distributions. E.g. in case of twinning, intensities are averaged and less reflections will have extreme (very high or very low) values and reflections will have more average values. I would expect that if you expand an untwinned P21212 data set to P21 and run the twin test, it still will come out untwinned. So if sfcheck claims the data is twinned, it most likely is and you have used the correct procedure. And in this case you should also process the data in the lower symmetry space group. What wonders me a little is that MR did not give clear solutions in P21 with twinned data. P21 is low symmetry and I would have expected 2 clear solutions, each corresponding to one of the two twin possibilities. Also, if the detwin procedure does what I think it does, it should not perform well if the twin fraction is very close to 0.5 (perfect twinning). As has been pointed out by others, you will have to generate the freeR flags in P21212 and expand them to P21, otherwise your free reflections are linked to working reflections and you will get artificially low free Rfactors. Also, Rfactors and free Rfactors will generally be lower if you use twin refinement. No reflections are discarded during twin refinement, what happens is that both twin orientations are taken into account during refinement (and for the calculation of Rfactors!). Congratulations, it looks like you have solved your problem! Herman ________________________________ From: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] On Behalf Of Andrey Nascimento Sent: Thursday, March 28, 2013 12:41 AM To: CCP4BB@JISCMAIL.AC.UK Subject: Re: [ccp4bb] Strange density in solvent channel and high Rfree Dear all, As I said in the latest topic, I could not model the third molecule. But when I superpose the two trimmers found in P1 MR solution (link below), I get the first two molecules perfect aligned and the third molecule inverted! (It is also possible to see the 2-fold axis and the third molecule lying on it!) I tried to run a MR with a model with two alternative positions and adjusted occupancy for the third molecule, but the Rfactor/free get higher (> 40%) and the map becomes worse - even the good ones (molecules 1 and 2) and for third molecule it remains bad (or worse). A procedure that "solved" the problem (decreased the Rfactor/free and gave good maps for third molecule) was the following: I integrated and scaled the data in P21, then I ran the sfcheck and it showed a twinned data (probably because of the (pseudo) higher symmetry present - P21212). So, I detwinned the data (with detwinn) and run a MR with detwinned data that gave a very good solution with tree molecules in ASU (it have never happened before!). After the MR I refined this MR solution against the original P21 data (without detwinn procedure) with amplitude based twin refinement in Refmac5 and, finally, it gave a good statistics (R factor / free about 0.19 / 0.22; FOM ~0.8) in the first round of refinement. I think that procedure probably discard reflections related to other positions making increasing the signal of the most frequent position. Link (.pdf): https://dl.dropbox.com/u/16221126/superposition.pdf Is there some problem in procedure described? If so, does anybody have a suggestion how can I model these disorder? Moreover, it seems to be a long range disorder (multiples positions along the all lattice), since even in P1 the maps for this third molecule are very bad. Thank you for all the suggestions. Cheers, Andrey 2013/3/25 Eleanor Dodson <eleanor.dod...@york.ac.uk<mailto:eleanor.dod...@york.ac.uk>> First - I dont think you have a 3rd molecule where you have put it - or at least not one with full occupancy. Those maps are a clear indication that something is wrong. What is the Matthews coefficient for the numbers in the asymmetric unit? Presumably your processing gave you a lattice which fitted the diffraction spots? ie you didnt miss a set of observations? You should see that at the data processing stage, and the different integration programs also try to report it. If there is non-crystallographic translation that can confuse things a bit; some classes of reflections might be systematically weak, but you can find if there is such a phenomena by doing a patterson. Or run ctruncate after merging the data - it checks this, and so does Xtriage. All these options will also check for twinning. If there is NCT then that could explain the high Rfactor. Are the spots nicely shaped? There are some cases of sheared crystals, which usually show up in distorted diffraction spots. If this is so and you have integrated the data according to an orthogonal lattice, there is nothing to stop you merging those observations in a low symmetry. Pointless gives you good statistics on the scoring for different symmetry operators. You can either run MR again in that symmetry - check all SGS consistent with the pointgroup, or try to work out how to position your P22121 solution in the new SG. There may well be 2n+1 copies of your molecule when you double the size of the asymmetric unit - all hard to check without more information. Good luck Eleanor On 22 March 2013 17:54, Andrey Nascimento <andreynascime...@gmail.com<mailto:andreynascime...@gmail.com>> wrote: Dear all, I have tried the procedure recommended by Zbyszek, expanding data from a higher symmetry and keeping the R-free set. But the map for third molecule (new molecule placed) are still very bad, even when a tried to reprocess data in P1 or P2 (P 1 21 1). The previous placed molecule (present in P2 21 21 ASU) and its symmetry related on P21 shows a very good map, but the third molecule are almost completely wrong (~50 residues in 470 are placed in quite good map) and map does not have connectivity to build a new molecule (even in lower sigmas, 0.8-1.0). I have tried automatic model building (AutoBuild and ARP/wARP) but they cannot build anything that make some sense or build a random chains without any sense. I do not have an extensive knowledge of crystallography, but I have been thinking about some questions: If the third molecule (the bad one) is lying on the 2-fold symmetry axis on P 2 21 21, and since it does not have an intrinsic 2-fold symmetry axis (like protein molecule), how can I merge the structure factors (or intensities) related by symmetry and expand to lower symmetry afterwards? In this case the molecule lying on the 2-fold symmetry axis will have the structure factors wrongly merged, since the molecule is not symmetric, is it ok? If the third molecule is lying on the 2-fold symmetry axis on P 2 21 21, and only another two molecules can be related by the crystallographic symmetry, is it a case of pseudo-symmetry? But in this case, the third molecule is disordered in the crystal packing (as Zbyszek said), and probably have a long range disorder, because I cannot get a good maps for this third molecule even in P1. (pseudo-symmetry + order/disorder????). And a more philosophical question... what is the problem in process data in a lower symmetry? Are there mathematical/statistical problems related that can lead to "false-good" data? I put a new .pdf file (ccp4bb_maps_P21.pdf) with map figures in this link: https://dl.dropbox.com/u/16221126/ccp4bb_maps_P21.pdf I am sorry for so many questions and thanks in advance. Cheers, Andrey 2013/3/20 Jrh <jrhelliw...@gmail.com<mailto:jrhelliw...@gmail.com>> Dear Zbyszek, I am concerned that the unmerged data would be bypassed and not preserved in your recommendation. I also find it counter intuitive that the merged data would then be unmerged into a lower symmetry and be better than the unmerged data; there is I imagine some useful reference or two you can direct me to that may well correct my lack of understanding. Thirdly I think this a very likely useful case to preserve the raw diffraction images. All best wishes, John Prof John R Helliwell DSc On 19 Mar 2013, at 14:37, Zbyszek Otwinowski <zbys...@work.swmed.edu<mailto:zbys...@work.swmed.edu>> wrote: > It is a clear-cut case of crystal packing disorder. The tell-tale sign is > that data can be merged in the higher-symmetry lattice, while the number > of molecules in the asymmetric unit (3 in P21) is not divisible by the > higher symmetry factor (2, by going from P21 to P21212). > From my experience, this is more likely a case of order-disorder than > merohedral twinning. The difference between these two is that structure > factors are added for the alternative conformations in the case of > order-disorder, while intensities (structure factors squared) are added in > the case of merohedral twinning. > > Now an important comment on how to proceed in the cases where data can be > merged in a higher symmetry, but the structure needs to be solved in a > lower symmetry due to a disorder. > > !Such data needs to be merged in the higher symmetry,assigned R-free flag, > and THEN expanded to the lower symmetry. Reprocessing the data in a lower > symmetry is an absolutely wrong procedure and it will artificially reduce > R-free, as the new R-free flags will not follow data symmetry! > > Moreover, while this one is likely to be a case of order-disorder, and > these are infrequent, reprocessing the data in a lower symmetry seems to > be frequently abused, essentially in order to reduce R-free. Generally, > when data CAN be merged in a higher symmetry, the only proper procedure in > going to a lower-symmetry structure is by expanding these higher-symmetry > data to a lower symmetry, and not by rescaling and merging the data in a > lower symmetry. > > Zbyszek Otwinowski > >> Dear all, >> We have solved the problem. Data processing in P1 looks better (six >> molecules in ASU), and Zanuda shows a P 1 21 1 symmetry (three molecules >> in >> ASU), Rfactor/Rfree drops to 0.20978/0.25719 in the first round >> of refinement (without put waters, ligands, etc.). >> >> Indeed, there were one more molecule in ASU, but the over-merged data in >> an orthorhombic lattice hid the correct solution. >> >> Thank you very much for all your suggestions, they were very important to >> solve this problem. >> >> Cheers, >> >> Andrey >> >> 2013/3/15 Andrey Nascimento >> <andreynascime...@gmail.com<mailto:andreynascime...@gmail.com>> >> >>> *Dear all,* >>> >>> *I have collected a good quality dataset of a protein with 64% of >>> solvent >>> in P 2 21 21 space group at 1.7A resolution with good statistical >>> parameters (values for last shell: Rmerge=0.202; I/Isig.=4.4; >>> Complet.=93% >>> Redun.=2.4, the overall values are better than last shell). The >>> structure >>> solution with molecular replacement goes well, the map quality at the >>> protein chain is very good, but in the final of refinement, after >>> addition >>> of a lot of waters and other solvent molecules, TLS refinement, etc. ... >>> the Rfree is a quite high yet, considering this resolution >>> (1.77A).(Rfree= >>> 0.29966 and Rfactor= 0.25534). Moreover, I reprocess the data in a lower >>> symmetry space group (P21), but I got the same problem, and I tried all >>> possible space groups for P222, but with other screw axis I can not even >>> solve the structure.* >>> >>> *A strange thing in the structure are the large solvent channels with a >>> lot of electron density positive peaks!? I usually did not see too many >>> peaks in the solvent channel like this. This peaks are the only reason >>> for >>> these high R's in refinement that I can find. But, why are there too >>> many >>> peaks in the solvent channel???* >>> >>> *I put a .pdf file (ccp4bb_maps.pdf) with some more information and map >>> figures in this link: https://dl.dropbox.com/u/16221126/ccp4bb_maps.pdf* >>> >>> * >>> * >>> >>> *Do someone have an explanation or solution for this?* >>> >>> * * >>> >>> *Cheers,* >>> >>> *Andrey* >>> >> > > > Zbyszek Otwinowski > UT Southwestern Medical Center at Dallas > 5323 Harry Hines Blvd. > Dallas, TX 75390-8816 > Tel. 214-645-6385<tel:214-645-6385> > Fax. 214-645-6353<tel:214-645-6353>
