On Tue, Jun 02, 2009 at 12:57:59PM -0400, Ed Pozharski wrote:
> I've seen this before in some structures and have verified in those
> cases that correction of the bulk solvent mask fixes the negative
> density. 
> ...
> It was quite convincing that it's indeed empty cavities large enough
> to be filled with artifact bulk solvent.

We do that by default in BUSTER now for ~5 years - also with good
experiences (all assuming the data isn't particularly weird at the low
resolution end - as discussed by others).
 
> In general, my experience was that although the negative density
> disappears, effect on R/Rfree is negligible (less than 0.1%).  They seem
> to be cavities made exclusively out of hydrophobic sidechains (i.e.
> carbon-only), but may set of observations is, of course, limited.

We see very rarely an increase in R/Rfree - but this nearly always
points back to very poor data and/or extremely partial models. In the
vast majority it doesn't hurt at all and in a lot of cases it gives
some significant drop in Rfree (0.5%) and/or reduction of Rfree-R
gap. But the nicest thing is: it makes looking at those Fo-Fc maps so
much easier - you don't get distracted by negative difference density
caused by wrong bulks solvent modeling.

Cheers

Clemens

-- 

***************************************************************
* Clemens Vonrhein, Ph.D.     vonrhein AT GlobalPhasing DOT com
*
*  Global Phasing Ltd.
*  Sheraton House, Castle Park 
*  Cambridge CB3 0AX, UK
*--------------------------------------------------------------
* BUSTER Development Group      (http://www.globalphasing.com)
***************************************************************

Reply via email to