On Mon, Apr 4, 2011 at 11:11 AM, Damian Ekiert
On a final note, regarding those pesky missing side chains: any thoughts on trying to employ a "Ringer"-like approach to model some of these (Fraser, et al., Nature 2009, 462(7273):669-673)? Is this practical (maybe this would add to many additional parameters)?
This is potentially useful for finding and building alternate conformers (something that most programs don't do - the only one I'm aware of is qFit: http://smb.slac.stanford.edu/qFitServer/qFit.jsp). But Ringer still relies on having some interpretable (albeit weak) density for the sidechains, and was designed to look at static rather than dynamic disorder. (A somewhat artificial distinction, but appropriate enough when talking about refinement.) It is also limited to relatively high resolution, usually better than 2.0A, not because of data-to-parameter ratio, but because the maps at lower resolutions just don't have enough detail to detect alternate conformations with any degree of confidence. Regarding missing or patchy domains, Pavel recently added a feature that should at least improve the phases and refinement behavior, but I'll let him describe it since I don't really understand what it does. I do not know of cases where people have found a reasonable way to model these explicitly, other than placing a rigid domain and letting the B-factors go crazy. -Nat