Hi Drew, Sorry for the slightly late reply. Something that 'might' be able to help (though I fully agree with Dale's reply) is the procedure for local density improvement described here:
Acta Cryst. (1997). D53, 540-543 "Local Improvement of Electron-Density Maps". We are currently in the process of implementing this in phenix. We have a preliminary command line tool (phenix.grow_density) ,and though it's not anywhere near ready for general use yet, if you want we could have a look at your data, as we are currently looking for test cases. Anybody else who feels they may benefit from this procedure and is willing to let us test it on their data is also welcome to contact me offline. Lester Carter PSI Technology Portal: http://technology.lbl.gov/portal/ LBNL , Berkeley X-RAL-MFrom: <aw...@med.nyu.edu> X-RAL-Connect: <ictmailer1.itd.rl.ac.uk [130.246.192.56]> Date: Wed, 20 May 2009 15:45:43 +0100 Reply-to: Drew Waight <aw...@med.nyu.edu> Sender: CCP4 bulletin board <CCP4BB@JISCMAIL.AC.UK> Hello all, I'm just finishing up my first structure, a membrane protein with 5 fold NCS. The native dataset is good to about 2.1A. The R factors are good, around .22/.26 without waters. The only problem is that I'm having a difficult time fitting what looks to be a very important (potential binding site) loop region of about 20 AA or so because the density is poor. I have relied heavily on the NCS to solve the rest of the structure, but releasing the constraints doesn't particularily help. I can see green blobs here and there in the difference map, and there is some continuous density at .8 sigma but its much poorer than the rest of the protein. (which is also mostly alpha helical) I would imagine that this is a common problem and am wondering if there are any tips from the experts seeing as I am extremely new to this. Thank you all, I have already learned a great deal from this board. Drew P.S. I have been using CCP4, Coot, Refmac, DM, Resolve (for dm), and Phenix refine
