On Tue, Feb 7, 2012 at 12:08 AM, Dialing Pretty
After we refine the structure of the protein to satisfactory with satisfactory Rwork and Rfree, we pick water by phenix refine, and I find Rfree always increases slightly after the water picking refinment.
Two thoughts: 1. If this is the same data you showed me before, the resolution is very marginal for automatic water picking - you're probably better off letting Coot find them and filtering them yourself (maybe after a round of refinement, or even just creating new maps). 2. In my experience, the water picking in phenix.refine works very well for medium-resolution structures (say 2.5-1.5A) when you're starting out with zero waters - it's great if you just want to improve the phases and maps quickly. It's less ideal for structures near the end of refinement where you already have a lot of waters placed. I've been running a lot of refinements of deposited structures recently, and there are almost always obvious water peaks that don't get filled (or rather, any waters that were there are removed) because they don't quite meet the relatively strict criteria. (There are also many peaks that aren't really waters, but that's a separate problem.) -Nat