Thank you for getting back so quickly.  

I am not starting out from scratch, per se, I have protein models based upon a 
threading analysis conducted on my protein's amino sequence vs a large fold 
database.  I did a psipred secondary structure prediction, used that as input 
to Threader 3.5, took the best matches threader produced and had threader 
output my protein as a pdb file based upon the best models it found.  The 
models are mostly OK...it is just that in a couple places, rather than 
maintain a fully contiguous strand of peptide in the model, threader has 
matched some short segments of my protein to the folding model template such 
that the backbone is broken and, in a couple cases, the ends are separated by 
quite a bit of space.  

Looking at the models it is easy to see that the end of beta strand X can 
easily (and must) connect to the beginning of beta strand Y...but for some 
reason, due to the software algorithm in threader, it elected not to do so.  
So what I have done is use the essentially "valid" model output by threader 
and connect the segments into a contiguous protein in pymol.  With some trial 
and error I have been able to get bond lengths to be correct and appropriate 
- mainly by moving residue a into close proximity to residue b and then 
building a bond between appropriate atoms.  

praedor

On Thursday 04 May 2006 12:58 pm, Warren DeLano wrote:
> Praedor,
>
> The sculpting capability in PyMOL is only intended and/or useful for
> handling conformational changes when starting from valid 3D geometries.
>
>
> If you are forming bonds or building molecules from scratch, then you'll
> need to use an external tool like OpenEye's Szybki or Schrodinger's
> MacroModel in order to get a valid starting model.
[...]f
> > Praedor Atrebates
> > Sent: Thursday, May 04, 2006 9:25 AM
> > To: [email protected]
> > Subject: [PyMOL] Correcting protiein models
> >
> > I have produced several threading models of a protein (the
> > app threader 3.5 can output pdb format models) I am studying.
> >  The models are OK for the most part but there are often odd
> > gaps in which the peptide backbone has simply been "broken"

-- 
The Reichstag fire is to Hitler as 9/11 is to Bush

Attachment: pgpCLy2IIjo8C.pgp
Description: PGP signature

Reply via email to