Hi Jeffrey, I would start this with phenix.xtriage. Is the twinning analysis (in particular the moments of the intensities) consistent with 50% twinning? It is not quite clear what you are describing with:
However, MolB (a 6 helix-bundle protein) is only half visible and the space of the invisible helices is occupied by the same visible helices
Perhaps a link to a picture would be helpful to visualize this. It would seem that you have a very low R/Rfree for having a significant part of the model still unmodelled. What happens if you leave out everything that is at all unclear. What is R/Rfree now? What does the map look like in the region in question? All the best, Tom T ________________________________________ From: [email protected] [[email protected]] on behalf of Luo, Jeffrey [JRDUS] [[email protected]] Sent: Tuesday, February 21, 2012 3:49 PM To: PHENIX user mailing list Subject: [phenixbb] Overlapping molecules Hi All, I have a situation that I would appreciate any/all help from the community. Crystal appears to be C2221 with a complex of MolA:MolB. Diffraction to ~2.2A with 1 complex in the ASU. Rmerge 5.7% to 2.2A R/Rfree=18.9/23.5% However, MolB (a 6 helix-bundle protein) is only half visible and the space of the invisible helices is occupied by the same visible helices of a crystallographic 2-fold (along a-axis) related MolB. This, of course, cannot be physically true. Two alternative trials. 1) indexed in P21 (same unit cell) with the 21 of C2221 as the unique b axis. Solution has 4 complexes. The c2221 a direction 2-fold should be absent in P21. However, the 4 independent complexes pack exactly as in C2221. 2) indexed in P1. 8 molecules. Packing is again identical to that of C2221. In both cases, R/Rfree are about the same. When used the MolA/MolB (full molecule as search model), then no solutions were found by Phaser due to severe clashes. My guess is that the crystal is P21 with beta=90 and there is twinning (2-fold along a-axis). And the twin fraction is 50%. At the same time, the missing helices must also be disordered in each twin fraction. Does this sound plausible? Is there a way to incorporate the twin law in MR to break this ambiguity? Thanks a lot! Jeffrey _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb