I agree that density looks dubious, the B-factor distribution among DNA atoms is odd (sudden shifts from blue to red) and the little wwPDB bar charts are on the red side.
But what about the density for the protein? The structure is of a hexameric helicase with with dsDNA going through the hole, so the space group being nearly-R3 or R32 doesn't shock me: one would expect the DNA to subtly break the symmetry. Could it simply be very poorly ordered DNA going making a pseudo-continuous helix going through the crystal? Or maybe the symmetry isn't broken, and DNA backbone is in 3 different rotational states even if there's rough density for the planes of the bases? ++++++++++++++++++++++++++++++++++++++++++ Phoebe A. Rice Dept. of Biochemistry & Molecular Biology The University of Chicago pr...@uchicago.edu https://voices.uchicago.edu/phoebericelab/ ________________________________________ From: CCP4 bulletin board [CCP4BB@JISCMAIL.AC.UK] on behalf of Anastassis Perrakis [a.perra...@nki.nl] Sent: Thursday, November 30, 2017 12:56 PM To: CCP4BB@JISCMAIL.AC.UK Subject: Re: [ccp4bb] Electron density in PDBe Interesting structure: a=b=c=α=β=γ=107 in P1 Estimated twinning fraction 0.439 for k,l,h and 0.439 for l,h,k In other words R32? ;-) The PDB-REDO map has some interesting features, one strand has bases density at .0 sigma, the other not, see attched. The NCS averaging they used is kind of interesting. It reads correct if I read through, but if something is done wrong, I am guessing it might result in erronous features. As a side note, the refereeing in eLife is interesting. The reviewing editor is Stephen Bell; an excellent scientist, but not a crystallographer. Referees opted to stay anonymous - but one could speculate they are not crystallogrpahers either, as they did not raise any crystallographic questions. The validation summary, well, it could have raised some red flags … sorry, it did raise red flags, or to be exact red bold boxes. Evidently, nobody looked at them. Have fun - A. [cid:B25D7E2E-FD64-40A6-9DBD-60DEF50D9249@home] [cid:5B0AA6A0-834B-46C7-AE56-8C6DFFB2F62D@home] A. On 30 Nov 2017, at 19:01, Wim Burmeister <wim.burmeis...@ibs.fr<mailto:wim.burmeis...@ibs.fr>> wrote: Dear all, I am doing some analysis on pdb entry 5tct. I realized that the electron density on the PDBe site (obtained from EDS) shows much more model bias than the one obtained from the pdb file and the cif reflection data file using ccp4i (import + refmac5 with 0 cycle positional refinement). Using the mtz file from pdb-redo a similar map is obtained. When I read the publication 5kleywegt et al, 2004) on EDS, I understood that essentially the same approach is used (refmac with 0 cycle refinement followed by fft map calculation). Does anybody have an idea where the discrepancy comes from ? I join a jpg with the electron density of the 3 approaches. The contour level is about 3 sigma. Best, Wim ps: I expect that the electron density of the DNA is not really present as the entry has a number of flaws. -- Wim Burmeister Professeur Institut de Biologie Structurale (IBS) CIBB 71 avenue des Martyrs CS 20192 38044 Grenoble Cedex 9, FRANCE E-mail: wim.burmeis...@ibs.fr<mailto:wim.burmeis...@ibs.fr> Tel: +33 (0) 457 42 87 41 Fax: +33 (0) 476 20 94 00 website<http://www.ibs.fr/research/research-groups/viral-replication-machines-group-m-jamin/team-03/article/poxvirus-replication-machinery-presentation/> <for-bb.jpg>