[Dock-fans] Fwd: Long bonds

Sat Oct 27 12:22:24 PDT 2007

Hi,

On Sat, 27 Oct 2007, Francesco Pietra wrote:

> --- Scott Brozell <sbrozell at scripps.edu> wrote:
> 
> > On Mon, 22 Oct 2007, Francesco Pietra wrote:
> > 
> > > Sorry, forgot: how to recognize where the protons are on histidine, so that
> > > HIE can be changed to HID? Similarly for aspartic.
> > 
> > I probably do not understand the question; did you examine the 
> > structures in a visualizer, such as Chimera, VMD, etc.
> > See the Amber score tutorial or the Amber manual for the definition
> > of HID, HIE, etc.
> > http://dock.compbio.ucsf.edu/DOCK_6/tutorials/amber_score/amber_score.htm
> > See below.
> 
> Thanks. Soon after posting, all that became clear using Chimera. 
> 
> As you kindly examined my problem, I feel myself obliged to tell what I also
> did in the meantime.
> 
> 
> > > --- Francesco Pietra <chiendarret at yahoo.com> wrote:
> > > 
> > > > With my large protein and ligand I was able to carry out docking, both
> > rigid
> > > > and flex ligand. Amber_score failed because of two kinds of problems with
> > the
> > > > protein.
> > 
> > > > 1st PROBLEM (protons on histidine, etc): 
> > > > FATAL:  Atom .R<GLU 443>.A<HE1 16> does not have a type.
> > > > FATAL:  Atom .R<HIE 441>.A<HD1 18> does not have a type.
> > > > FATAL:  Atom .R<ASP 402>.A<HD1 13> does not have a type.
> > > > FATAL:  Atom .R<ASP 386>.A<HD1 13> does not have a type.
> > 
> > > > Is that better in such cases to edit the names by hand (problematic with
> > > > aspartic, H on O), or make recourse to (a) downlodable sourceforge softw,
> > (b)
> > > > "biophysics" softw on server?
> > 
> > The appropriate method depends on several factors including your
> > confidence in the existing protonation versus your confidence
> > in those produced by tools, as well as whether the existing proton
> > states/coordinates need to be preserved.  See these for additional comments:
> > http://blur.compbio.ucsf.edu/pipermail/dock-fans/2007-January/000892.html
> > http://blur.compbio.ucsf.edu/pipermail/dock-fans/2007-January/000876.html
> 
> See PS, please. I'll go though these suggestions.
> > 
> > > > ----------------------------
> > 
> > > > 2nd PROBLEM:
> > 
> > > > WARNING: There is a bond of 50.040630 angstroms between: 
> > > > -------  .R<THR 333>.A<C 13> and .R<SER 334>.A<N 1>
> > > > WARNING: There is a bond of 39.464740 angstroms between: 
> > > > -------  .R<THR 222>.A<C 13> and .R<SER 223>.A<N 1>
> > > > WARNING: There is a bond of 49.372740 angstroms between: 
> > 
> > > > I was advised to insert "TER" records between the residues that the bond
> > is
> > > > connecting. For the first long bond, is that correct as follows:
> > 
> > The central issue is whether there should be a bond between these
> > residues ?  If not then insert TER cards after each separate
> > molecule.  If there should be a bond then the initial xyz coordinates
> > are screwed up.
> 
> 
> In removing peripheral helices from the protein, perhaps because of my modest
> experience, those long bonds formed from unsaturated valencies. Because they
> should not be present  (they just cross the opening of the pore in its widest
> "diameter", as detected with Chimera, while VMD - as I used it  default - did
> not reveal them) I inserted "TER" on columns 1-3 between the relevant residues,
> whereby there were no more such long bonds on loading the pdb to Chimera. That
> occurred several days ago.
> 
> I did also other editing to pdb
> (a) rename HIS -> HID
> (b) Rename ASP and GLU -> ASH and GLH, resp. This required also exchanging the
> name of the oxygens and rename the proton, which was on the "wrong" oxygen.
> (c) Delete unwanted N-terminal 'H' hydrogens.
> 
> That cleaned much the protein, though not completely. Dock-Prep - through
> out-of-place questions, arrived at non-integral charges. At this point I was
> lost. The problem is being kindly examined by the Chimera team. I did not try
> to fix with Amber 9 tleap.

In most visualization programs one can select and remove pieces,
such as peripheral helices, and I would expect the result to be clean
with respect to capping dangling bonds.
Significant reworking of the starting structure is occurring.
In the big picture this may be reasonable, but getting all the details
correct may be challenging.  I recommend running some intermediate steps
through LEaP since it can be fussy.

I have not used pdb2pqr.

Scott

> PS: I also tried to use pdq2pqr. On forming the pqr file with the sole
> "--ff=amber" all three "TER" were removed without compensating with hydrogens.
> The file did no more open with Chimera, which found alleged non-adherence to
> the pdb format. VDM opened it. The option "--with-ph=7.4" (which is just the pH
> at which experiments have been carried out with the protein) the error was
> "unable to find propka", although the compilation with propka was correct. I
> dropped the matter because that probably is not the right way to get the pdb
> following the rules.
> 
> 
> > 
> > Scott
> > 
> > ...
> > > > ATOM   5150  OG1 THR X 333      13.670 -12.723  22.556  0.00  0.00       
> >    
> > > > ATOM   5151  HG1 THR X 333      12.831 -13.194  22.522  0.00  0.00       
> >    
> > > > ATOM   5152  C   THR X 333      14.755 -14.634  24.443  0.00  0.00       
> >    
> > > > ATOM   5153  O   THR X 333      15.694 -15.150  25.132  0.00  0.00 
> > > > TER           
> > > > ATOM   5154  N   SER X 334     -16.176  22.312  10.940  0.00  0.00       
> >    
> > > > ATOM   5155  H   SER X 334     -15.301  22.760  10.966  0.00  0.00       
> >    
> > > > ATOM   5156  CA  SER X 334     -16.224  20.932  11.429  0.00  0.00       
> >    
> > ...
> > 
> > 
> 
> 
> __________________________________________________
> Do You Yahoo!?
> Tired of spam?  Yahoo! Mail has the best spam protection around 
> http://mail.yahoo.com 
> 

-- 