On Fri, Oct 22, 2010 at 11:01 AM, Ed Pozharski <epozh...@umaryland.edu>wrote:
> Ethan pointed out to me that Babinet-principle based solvent correction > is not always inferior to mask-based approach. On Friday, October 22, 2010 11:17:53 am Pavel Afonine wrote: > Hi Ed, > > may be "not always" (in a sense that mask-based one sometimes may leave a > negative footprint in areas where there is no solvent), but I still want to > remind this paper where there is some discussion about exponential vs > mask-based bulk solvent model: > > http://journals.iucr.org/d/issues/2002/09/00/fw0020/index.html > > Also, if you refer to "some old observations" then I believe since that the > bulk solvent correction protocols have been improved very significantly: > > http://journals.iucr.org/d/issues/2005/07/00/gx5040/index.html > and > A.T. Brunger. Nature Protocols 2, 2728-2733 (2007). Pavel: Thank you for the pointer to those papers. Unfortunately for the current discussion, I do not find any practical comparison to Babinet models in either paper. Both start with the implicit assumption that a flat, masked model is better and then proceed to explore how best to determine its parameters. The Fokine and Urzhumtsev paper does state that "the [Babinet] exponential scaling model is handicapped at higher resolution", but does not present any examples or comparisons to document what effect this has in practice. The same is true for the very recent paper by Fenn, Schneiders, and Brunger Acta Cryst. (2010). D66, 1024-1031 [ doi:10.1107/S0907444910031045 ] This paper presents the theoretical basis for the Babinet treatment and for several recent hybrid mask treatments that get away from describing the solvent region as "flat". But the paper does not include a Babinet model in the tables of comparative results. Do you know of any published or unpublished results that compare the R factors achieved by Babinet treatment with those obtained from the state-of-the-art mask models? I'll cc this to Tim Fenn in case he wants to chime in with additional data that wasn't included in the recent paper. All I can say is that based on my own experience in refining structures at various resolutions, a Babinet model often, but not always, yields better R and Rfree than a masked solvent model. This held true when I was using X-PLOR (later CNS) and still holds true today using refmac. Shelxl also uses a Babinet model, which is relevant because shelxl is generally used for the highest resolution refinements, even though this is exactly the regime in which the Babinet model is argued to be weakest. I have not done side-by-side comparisons in phenix, nor have I tried the hybrid models described by Fenn et al. My advice to Ed, which I now repeat here, was simply to try both refmac options in parallel. It only takes a few minutes! I know of no way to predict in advance which option will yield better R factors for your particular structure, though I suspect there are some properties of the data or of the model that are relevant. happy modeling, Ethan -- Ethan A Merritt Biomolecular Structure Center, K-428 Health Sciences Bldg University of Washington, Seattle 98195-7742
