Hi Wei, thanks for sharing the data (off-list). I did some detective work and yes, Clemens is correct: what you see is the effect of bulk-solvent. After adjusting the solvent contribution (mask) in regions occupied by altlocs A and B the positive density mostly disappears.
Pavel On Fri, Jul 21, 2017 at 12:26 AM, Clemens Vonrhein < vonrh...@globalphasing.com> wrote: > Dear Wei, > > remember that you might have "something else" in the location occupied > by each alternate conformation when it is not occupied by that > particular conformation. If you only model two alternate conformations > you are saying something like > > that space A is occupied 50% of the time by this A conformation and > 50% it is vacuum > > that space B is occupied 50% of the time by this B conformation and > 50% it is vacuum > > It is more likely that you will have space A occupied N% by altConf A > and (100-N)% e.g. by water/solvent. And for B you have (100-N)% > altConf B and N% e.g. water/solvent. So you will need to model > everything marked as N% as altConf A and everything marked (100-N)% as > altConf B - including water/solvent. > > Because this is not yet done, the occupancy refinement might try to > explain the non-vacuum density (water/solvent) by increasing the > occupancy of each conformation so they sum up to larger than 1. > > Does that make sense? > > Cheers > > Clemens > > PS: in BUSTER (http://www.globalphasing.com/buster/) you can restrain > the summed occupancy to one, which can help clarifying how to > model the water/solvent region underneath each conformation > ... depending on data quality/resolution of course. I'm sure other > refinement programs have a similar feature. > > PPS: it can be sometimes useful to start refinement once from > A=0.1/B=0.9 and a second run with A=0.9/B=0.1. Refinement should > reach very similar final values in both cases - which is also a > good way of providing confidence in the final results, I think. > > > On Fri, Jul 21, 2017 at 01:10:15AM +0000, Wang, Wei wrote: > > Hi Everyone, > > > > > > One quick question: > > > > > > When I do phenix refinement, I see continuous omitted map at one RNA > chain 3'-end (Fig1). > > > > There is no possibility of next nucleotide because I have built > full-length RNA sequence, which is well fitted into density. > > > > Hence two conformations is most possible for the last nucleotide. And > our biochemical data also supported the result of two conformations. > > > > However, I found the occupancy is always > 1.0 (fig2 and fig3), when I > did refinement of occupancy using phenix refinement,. > > > > I also think about the possibility of small molecules in the crystal > buffer. But in most structures of this protein, there is always no small > molecule here. > > > > > > Any suggestions for occupancy refinement? I will appreciate your great > help. > > > > > > Best, > > > > Wei > > > > > > > > [cid:2b30f1a8-2a3a-4483-ac12-310f4bbed70b] > > > > > > > > > > > > > > > > -- > > *-------------------------------------------------------------- > * Clemens Vonrhein, Ph.D. vonrhein AT GlobalPhasing DOT com > * Global Phasing Ltd., Sheraton House, Castle Park > * Cambridge CB3 0AX, UK www.globalphasing.com > *-------------------------------------------------------------- >
