On Mon, Dec 12, 2011 at 8:26 PM, intekhab alam
I have a 3.0A dataset of a protein-protein complex. I used one of the protein structure solved previously as a template and used phaser in CCP4 as well as in phenix. I used the sca file in phenix which gave a solution with 6 monomers in ASU which are packed as a hexamer. In ccp4 phaser i used the converted .mtz file which gave 6 molecules in ASU but with a different packing. In both the cases map fits well in the model and there is some extra density that may correspond to second protein partner of the complex molecule. when i tried to use the MR solution generated in phenix with the mtz file from ccp4 the model shows a lot of clashes.
The two solutions are probably equivalent under crystallographic symmetry - they simply represent different versions of the ASU. It's possible that they also have different origins, but I can't tell from your description whether this is the case. What kind of refinement are you running that results in clashes, and what is the clashscore for the resulting model? (Please include the program version.) As a general rule, reciprocal-space refinement without phases should work regardless of the origin or ASU, so it could just be a weighting problem. If you are running real-space refinement that is potentially more problematic, but I doubt that's the problem in this case. -Nat