Hi James, This kind of calculation is relatively painful compared to a monomeric structure, so it'll probably take some toying around to make it work. We don't have a script that just works for this kind of thing, unfortunately.
The COLLapse term (Rgyr restraint) can have multiple independent assignments--just repeat the COLLapse ASSIgn statement with different selections. That's probably worth doing, rather than relying on the symmetry restraints. And of course, you want to enforce the symmetry with the non- crystallographic symmetry term, not the crystallographic symmetry operations. To keep the dimers magnetically equivalent to each other, you might need to use Michael Nilges's distance symmetry NOE restraints. He has a script for generating them on his web site, I believe. Most likely, the best way to proceed would be to generate coordinates for the dimer, and then use GMC's rigid body approach. Hope this helps. --JK On Oct 18, 2007, at 1:57 AM, James wrote: > Excuse this if it has arrived twice. > > Greetings > > Just done an NMR structure of a protein that can form a hexamer. > I'd basically like to model/generate the hexamer using the NMR > structure of the monomer. > There is a protein with good sequence identity (45%) that is a > heptamer, to model the likely interfaces. > > I think it may be possible to use Marius' RGyr term/Rigid body > docking methods to assemble the hexamer using partial ridgid > bodies, floppy sidechains etc > and to include a couple of loose synthetic NOEs) along with > crystalography symmetry terms. > > Is the best way to do this to first build up 3 dimers using the > RGyr/NOEs (they will not be symmetric, head-tail) > then asemble the hexamer from dimer1->dimer2 then dimer1-dimer-2- > >dimer3?? > > But can I define several different radius of gyration terms? or 3 > or even 6 and use crystalographic symmetry. i.e. do the lot at once? > > Has anybody done this kinda thing? Is there an obviously easy easy > way that I've missed? > > I can model a hexamer rather than the heptamer on something else > but the interface may not be as good as the sequence identity is > much lower. > then I can do the 6 RGy + some NOE from the heptamer etc.... > > thanks > James > > _______________________________________________ > Xplor-nih mailing list > Xplor-nih at nmr.cit.nih.gov > http://dcb.cit.nih.gov/mailman/listinfo/xplor-nih
