Drop Your Rfree and Rwork by 1 %
I have tried this method on 4 different structures and so far it has dropped both the Rfree and Rwork by about 1 percent, but your mileage will vary. For better or worse, one can end up spending a great deal of time trying to lower their R values. I wanted to share with you the steps of how I have been able to drop these values.
1) Complete your structure
-have it to the point that you would submit it to the PDB (maintain your Rfree)
2) Go back to the point in refinement that you were satisfied with the protein structure before adding solvent
3) Add all your solvent from your final structure
-paste them into the pdb file
4) Refine as normal
Your mileage will vary based on:
Crystal packing
Resolution
[add your own here]
Give this a try and drop me a line with your results.
Sili
October 19th, 2009 at 4:05 AM #
Any idea why that gets you out of the false minimum? Or why you end down there in the first place?
Of course, I’ve never done protein work, so I may need to dummies’ explanation.
Sean
October 19th, 2009 at 1:27 PM #
I add the solvent in stages using ARP/WARP. I believe this is what creates a false minimum.
The protein shifts ever so slightly to make up for the missing water. I think by adding all the solvent at once it allows the structure to reach a lower minimum.