Dear Wulf, I have had a similar experience recently when water treatment in SHELXL became a nightmare - as you say they started shifting away even if those were clear water sites. I have not found a final solution to this but I was not worried by those where density "disappeared" - those were the weak ones anyway so I attributed this behaviour to the different refinement algorithm and treatment of bulk solvent etc.
I still believe that this is related to the max. likelihood vs. conjugated gradients and FFT vs. full summation Fourier and bulk solvent treatment in those programs. My way was to leave this water fiddling to the very last cycle and not let them go away too much. Those that went completely wrong were deleted. I wish I had a more satisfying answer. Jan On Thu, Nov 12, 2009 at 8:53 AM, Wulf Blankenfeldt < wulf.blankenfe...@mpi-dortmund.mpg.de> wrote: > Dear all, > > I am blessed/tormented with a 0.8 A data set for a small protein. I first > refined it using COOT/REFMAC5 until I deemed it done, which meant manually > adding quite a few water molecules that were weak and a little more distant > to the protein. R is 10.2%, R-free is 10.8%. > > I realize that referees will probably ask for a SHELXL refinement, so I > spent the last few days learning about FVAR and SUMP instructions and the > like and finally got it running in a way I thought should be correct. Using > my final REFMAC5 model, I get R=10.8% and R-free=11.1% - I can certainly > live with this, but: > > When I now look at the electron density, a large number of my tenderly > curated water molecules have lost their 2FO-FC electron density (50 out of > 200; I usually use a 1 sigma cutoff), or they have shifted to lie beside a > density blob. It may have to do with the fact that I assigned 0.5 occupancy > to some of them and now they have been BUMPed away, others may simply be too > weak to stay in place. > > I wonder what to do now. Do I have to rebuild the water structure again and > refine everything in SHELXL exclusively? Or are there any flags I could use > to keep the old waters intact (risking higher R-factors)? > > Any help is highly appreciated! > > Thank you in advance, > > > Wulf > -- Jan Dohnalek, Ph.D Institute of Macromolecular Chemistry Academy of Sciences of the Czech Republic Heyrovskeho nam. 2 16206 Praha 6 Czech Republic Tel: +420 296 809 390 Fax: +420 296 809 410
