AMBER Archive (2004)

Subject: Re: AMBER: Defining residues

From: John Bushnell (bushnell_at_chem.ucsb.edu)
Date: Mon Feb 23 2004 - 17:39:46 CST

On Mon, 23 Feb 2004, Bill Ross wrote:

>
> > set cen tail cen.1.x (where x is the tail atom number)
> > tmp1 = sequence { cen arm1 }
> > set tmp1 tail tmp1.1.y (where y is where I want to attach arm2)
> > tmp2 = sequence { tmp1 arm2 }
> > set tmp2 tail tmp2.1.z (z is the new tail on the central residue)
> > tmp3 = sequence { tmp2 arm3 }
>
> Have you verified that the parts are actually bonded?

I make it a habit to always look at a structure before running it.
It's too easy to make a stupid mistake and then waste time running
junk (garbage in, garbage out).

As for the steric overlap problem, this can definitely be a problem,
and I have avoided it by using a geometry for my sidechains that wouldn't
interfere with the others. This allowed me to run hundreds of isomers
in production runs without any bad contacts.
 
> Using 'sequence' should guarantee that leap will try to orient
> the joined residues according to internal coordinates derived
> from the residue templates (i.e. trying to keep atoms being
> bonded oriented properly for ther hybridization), however there
> is no guarantee that steric overlap will be avoided.
>
> Another approach is to put together a pdb file with all parts aligned
> with TER cards between them using a more powerful graphical editor
> than leap, then loadpdb and use the 'bond' command in leap to join them.

Ah! Now I see how this could work. This could be very useful, though
I'm not sure offhand how a PDB created by some other package would be
recognized when made up of a bunch of custom residues created in Amber.

       - John

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu