Dear Andrew, This is a great detective work of yours. Hoewver, since these are data recorded at home, the obvious way out of this mess is to screen/collect data at a synchrotron (as suggeted by Andrew Thompson) with a much smaller oscillation range as pointed out by Phil. The much smaller irradiated crystal volumes may enable to deal better with this nasty mixture of lattices.
This is of course assuming that a search for different crystallization conditions has been exhausted. Cheers, Boaz Boaz Shaanan, Ph.D. Dept. of Life Sciences Ben Gurion University Beer Sheva, Israel On Aug 1, 2022 18:24, Andrew Leslie - MRC LMB <[email protected]> wrote: Dear Sayan, I believe that I have found the answer to your problem. When I tried indexing using images 1 & 2 (90° apart in phi) I could only find an orthorhombic solution, there was no suggestion of a C2 solution. As Eleanor, Land and myself have already pointed out, the P222 cell and the C2 cell that you provided in your first email cannot be interconverted, so there was obviously an issue because you were able to integrate the C2 cell giving a MR solution and a model with similar R-factors. When I tried the multiple lattice indexing in imosflm with 6 images, phi values of 0,30,60,90,120 and 150 it immediately found two lattices, the C2 cell and the P222 cell that you described. Thus your crystal is actually made up of two distinct components with different spacegroups. I believe that Zbyszek Dauter has published a similar example. The problem then is that many of the spots from the two lattices overlap on the detector, so the measured intensity will not be correct for either lattice. This will account for the high R-factors that you observe. Unfortunately, the strength of the diffraction from the two different lattices can vary as the crystal rotates, depending on the volume of each lattice that is in the beam. This is why indexing with just 2 images, at 0 and 90°, gives no indication of the C2 cell. I do not believe that there is any integration package that will allow you to “deconvolute” the intensities from the two different lattices where they overlap. imosflm gives you the opportunity to integrate one lattice and ignore any spots in that lattice that are overlapped by the second lattice, but this can lead to quite incomplete data. Often the best way to deal with this situation is to ignore one of the lattices and just integrate the other, and if the volume of that lattice in the beam is greater (ie the spots are stronger) this can give a good model. In your case the diffraction from the P222 lattice seems to be stronger (gives lower R-factors for the refined model). You can check the quality of the model by looking at the density for atoms that are not included in the model (eg the ligand), or by deleting some residues and seeing how well the density comes back. Best wishes, Andrew On 29 Jul 2022, at 12:17, Sayan Saha <[email protected]<mailto:[email protected]>> wrote: Dear Sir, The Rw/Rf for P22121 structure solution is ~29/32%. For C2 structure solution, it is a little higher, 32/35%. With best regards, Sayan Saha. On Fri, Jul 29, 2022 at 3:25 PM Andrew Leslie - MRC LMB <[email protected]<mailto:[email protected]>> wrote: Dear Sayan, Using imosflm, based on the two images that you have uploaded, the cell appears to be orthorhombic (approx 80, 85, 111) and there is no evidence for the C2 unit call that you suggested. Using only the second image there is a C2 solution, but the prediction is very poor (high positional error). I am therefore a bit surprised that you found a MR solution in C2. Were the Rfactors that you quoted (29/32%) for the orthorhombic solution or the C2 solution? As Herman pointed out, there is definite streaking in some lunes on image 2, but this seems to be restricted to a relatively small part of the diffraction pattern. While this does indicate some kind of disorder, I do not think this is serious enough to prevent a reliable structure determination, but it might account for the slightly high R-factors. There is definitely a lot of spot overlap, as the mosaic spread (mosflm definition) is in the region of 1.5°. The oscillation angle would have to be 0.3° or less to avoid this spot overlap (determined from the Strategy option in imosflm). Again, as Kay pointed out, this would lead to higher then expected R-factors. As this is an image plate detector, I can understand why you might not be using an oscillation angle of 0.1°, but you do need to check that the oscillation angle you are using does not give rise to a lot of spatial overlaps and a smaller oscillation angle will generally give improved quality data, especially if the background level is quite high, as it is in your images. Best wishes, Andrew On 29 Jul 2022, at 05:06, Sayan Saha <[email protected]<mailto:[email protected]>> wrote: Dear Sir, [https://ssl.gstatic.com/docs/doclist/images/icon_10_generic_list.png] image1.osc<https://drive.google.com/file/d/1K5hhoMymVyidOjZyR5Hbfb8ny-KkMIm6/view?usp=drive_web> [https://ssl.gstatic.com/docs/doclist/images/icon_10_generic_list.png] image2.osc<https://drive.google.com/file/d/16TsGwBPtrkVxOYN7M5PxvJyS0pvMljVZ/view?usp=drive_web> The detector-to-crystal distance was 190 mm. The Oscillation range was 1.0 degree. Please find attached two diffraction images. With best regards, Sayan Saha. On Thu, Jul 28, 2022 at 9:41 PM Sayan Saha <[email protected]<mailto:[email protected]>> wrote: Dear Sir, The detector-to-crystal distance was 190 mm. The Oscillation range was 1.0 degree. Please find attached two diffraction images. With best regards, Sayan Saha. On Thu, Jul 28, 2022 at 7:31 PM Kay Diederichs <[email protected]<mailto:[email protected]>> wrote: Dear Sayan, On Thu, 28 Jul 2022 15:12:30 +0530, Sayan Saha <[email protected]<mailto:[email protected]>> wrote: >Dear Sir, > >1. There are no ice-rings. However, diffraction spots seem to be >overlapping. This can be seen during the data processing, as the space >group (C2 or P222) varies even in the consecutive frames. spot overlap results in inaccurate intensity values. Inaccurate intensities result in high Rwork/Rfree. Why do the spots overlap? High mosaicity? Detector distance too small? Oscillation range too high (0.1° is typically adequate)? It would be good to see the data, otherwise we can only speculate. Space group does not change from one frame to the next. If you use XDS, a good guide to decide between higher and lower-symmetry space groups is to compare their ISa values. best, Kay > >2. Crystal packing of C2 and P22121 seem to be similar (please see the >attached images). > >3. Forgot to mention in my previous email that we have already processed >the data in P1 and MR solution could be found only in P1 (Phaser was used >with an option in all possible space groups of that point group). > >Please let me know if any other information is required. > >With best regards, >Sayan Saha. > > >On Thu, Jul 28, 2022 at 1:26 PM Schreuder, Herman /DE < >[email protected]<mailto:[email protected]>> wrote: > >> Dear Sayan, >> >> >> >> If a subunit is correctly oriented, but the translation is incorrect, >> density for a ligand may still show up in the binding site of the protein. >> It might be that one of the 2-fold axes, you think is crystallographic, is >> in fact non crystallographic and a few Angstroms away from the >> crystallographic position. >> >> >> >> What I would do: >> >> 1. Check the images: are there ice-rings or other artifacts that could >> cause scaling problems that would lead to high Rw/Rf values? In that case, >> there is not much you can do. >> 2. Compare the C2 and P22121 solutions: do they have the same overall >> crystal packing (CS+NCS), or are they different? Do they have the same >> Rw/Rf values? Can we learn anything from the differences in overall >> crystal >> packing? >> 3. Process, run MR and refine in P1. Do you get lower R-factors? If >> so, then run Zanuda to find out the real space group. >> >> >> >> Best, >> >> Herman >> >> >> >> *Von:* CCP4 bulletin board >> <[email protected]<mailto:[email protected]>> *Im Auftrag von *Sayan >> Saha >> *Gesendet:* Donnerstag, 28. Juli 2022 08:15 >> *An:* [email protected]<mailto:[email protected]> >> *Betreff:* [ccp4bb] Regarding the correct space group identification >> >> >> >> Dear All, >> >> >> >> We have collected home-source X-ray intensity data for a protein at 2.6 >> Angstrom. The data can be processed in either C2 (a=120, b=80, c=85 and >> beta=115) or P222 (P22121, a=80, b=85, c=110). MR solution can be obtained >> in both the space groups. However, the solution can be refined with an >> Rw/Rf of 29/32% only. The protein is bound to a ligand (co-crystallization) >> for which a clear density can be observed. >> >> >> >> Any help and suggestion in this regard would be very helpful. >> >> >> >> With best regards, >> >> Sayan Saha. >> >> >> >> >> ------------------------------ >> >> To unsubscribe from the CCP4BB list, click the following link: >> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB&A=1 >> > >######################################################################## > >To unsubscribe from the CCP4BB list, click the following link: >https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB&A=1 > >This message was issued to members of >www.jiscmail.ac.uk/CCP4BB<http://www.jiscmail.ac.uk/CCP4BB>, a mailing list >hosted by www.jiscmail.ac.uk<http://www.jiscmail.ac.uk/>, terms & conditions >are available at https://www.jiscmail.ac.uk/policyandsecurity/ > ######################################################################## To unsubscribe from the CCP4BB list, click the following link: https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB&A=1 This message was issued to members of www.jiscmail.ac.uk/CCP4BB<http://www.jiscmail.ac.uk/CCP4BB>, a mailing list hosted by www.jiscmail.ac.uk<http://www.jiscmail.ac.uk/>, terms & conditions are available at https://www.jiscmail.ac.uk/policyandsecurity/ ________________________________ To unsubscribe from the CCP4BB list, click the following link: https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB&A=1 ________________________________ To unsubscribe from the CCP4BB list, click the following link: https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB&A=1 ######################################################################## To unsubscribe from the CCP4BB list, click the following link: https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB&A=1 This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list hosted by www.jiscmail.ac.uk, terms & conditions are available at https://www.jiscmail.ac.uk/policyandsecurity/
