Hi Charles, On Thu, 15 Dec 2005 [email protected] wrote:
> -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > > Hi Greg-- > > > I've only recently begun using xplor, so I hope this isn't a stupid > > question. I am trying to refine the structure of my protein, which > > includes a covalently bound acyl phosphopantethein species (about 32 > > heavy atoms, and 20 protons. I have a good-looking structure for the > > protein, and I want to use xplor to add on the covalently attached > > species. I was able to use the addUnknownAtoms method to help out with > > this, though I notice that the structure I get out has a large number of > > bond, angle, and improper violations. Luckily, this doesn't seem to > > matter too much, as subsequent refinement seems to clean up the structure. > > > > a quick call to protocol.fixupCovalentGeom() should also clean things > up. Perhaps it should be called as the final step of addUnknownAtoms... Thanks very much. That does work quite well. However, there is a slight problem with fixupCovalentGeom() as it stands. If you only run it on a portion of a covalent structure, it will violate restraints between the selected and unselected part of the molecule. I've attached the fix as a patch - it changes the constraint definition to allow constraints between the selected and unselected parts of a molecule. The only problem with sticking fixupCovalentgeom into addUnknownAtoms is that you would probably have to export the fixupCovalentgeom parameters to the function call for addUnknownAtoms - I had to change maxIters to get fixupCovalentgeom to get the structure right. It also may be worthwhile to mention in the documentation that fixupCovalentgeom changes the 'constraints' setting, depending on what the user wants to do afterwards. > > > Once the group is attached, I wanted to refine its structure based on > > some ambiguous NOE restraints. I've tried using the scripts in > > eginput/prot_prot as a model, but I've essentially removed the 'fix' > > statements that held portions of the structure rigid. I also found that > > the rigid body docking step (rigid_min.inp) wasn't really necessary. (I > > had modified this step to allow the attachment to be fully flexible.) > > > > Couple of questions: > > Is there a better script to use as a starting point? > > What you're doing is probably fine, but why not use one of the Python > examples? > Is yours a docking problem? The main reason was because I have had a little bit more exposure to the xplor interface and am not as sure of things with the python interface. My problem isn't truly a docking problem, but I'm treating it similarly to a docking problem. I have a protein structure that I produced in cyana, along with distance and dihedral angle restraints from cyana that are now in xplor format. The structural data was acquired on a sample with a covalently bound small molecule, but the structure from cyana is only protein. There are a couple of things that I want to do with xplor: 1. Generate pdb files of the previous protein structure with the small molecule covalently attached. The small molecule in a random conformation, but with bonds, angles, impropers satisfied. I've now gotten this to work, with the help of fixupcovalentgeom(). 2. Insert the small molecule into the protein structure based on some ambiguous NOE restraints that are in a tbl file. 3. Refine that structure using the database potentials. (I've been told that these are as good, if not better than water refinement. #2 can be framed similarly to a docking problem, although I don't have to worry about collapse terms. Now that I'm thinking about it, I should probably be fixing my protein structure and initially shutting off VDW for the attached small molecule, gradually bringing VDW back as I lower the temperature. I guess once the small molecule is nestled in its binding pocket (yes, there is a binding pocket - I just need to get the small molecule in there.), the database potential refinement shouldn't be difficult. Comments and suggestions are welcome. Thanks, -Greg [email protected] -------------- next part -------------- --- python/protocol.py.bak 2005-12-09 12:12:43.000000000 -0600 +++ python/protocol.py 2005-12-16 10:25:31.000000000 -0600 @@ -565,9 +565,11 @@ xplor.command("""constraints fix (not (%s)) + interaction (%s) (not (%s)) interaction (%s) (%s) end""" % - (sel.string(),sel.string(),sel.string())) + (sel.string(),sel.string(),sel.string(), + sel.string(),sel.string())) xplor.command("flags exclude * include scripting end")
