Dear Thomas, There are many options to test before a diagnosis can be made.
When do you say that the electron density is bad, do you mean that it looks like a map for a lower resolution dataset or do you see other problems? What is the analysis from TRUNCATE? Do you see any anormality (i.e. L- and H- tests)? Firstly I would run Zanuda, which is part of the CCP4, using the .mtz and .pdb files. This is to check the SG choice. Secondly, are you refining the data using REFMAC with the twinning option active? Although this option is recommended to be used only in the latest stages of the refinement, it may give you a clue of the twinning law, if any. If you have enough data, I would integrate them in P1 and analyse them with AIMLESS. It should be always possible to trade crystal symmetry for NCS and you may even be lucky to solve the structure in P1, if there is enough signal. If you used MOLREP, it should tell you whether pseudo-translation is detected. You can also inspect the diffraction pattern symmetry with the program HKLVIEW, which is supported by CCP4 or with LABELIT (i.e. pseudo-precession camera images). In my experience MR solutions are possible even when twinning or other lattice defects remain unnoticed. The refinement is indeed the best of all judges. Good luck! Dr. Ivan Campeotto Imperial College London On 17 Apr 2014, at 16:18, Thomas RORET wrote: Dear all, one of my colleagues has probably a twin and he can't find the good solution or the twin law. Using XDS he has obtained the space group P212121 and a unit cell of a= 55.54 b= 60.72 c= 121.4 and all angles are 90°. In this unit cell b is approximately equal to 1/2 c. Datas are good for 1.9A resolution and a solution can be found by molecular replacement. Peaks of Rotation Function (Rf/sig) 1- 11.48 2- 6.80 3- 4.25 ... Translation function (Tf/sig) 1- 8.13 2- 5.21 3- 2.70 ... It is possible to make refinement but the electron density is not good and it is not possible to decrease the Rfree under 33.00%. If somebody can help me ... Regards -- Thomas RORET BioMod Team Tel. 00 333 83 68 47 89 CRM2 UMR CNRS-UL 7036 Faculté des Sciences et Technologies BP 70239 54506 Vandoeuvre-les-Nancy
