Date: Tue, 3 Feb 2004 09:44:50 -0800
From: "David A. Case"
Subject: Re: AMBER: leap - orientation of sequenced residues

On Mon, Feb 02, 2004, Sarah Wittkopp wrote:
>
> I am trying to determine where in the leap source code the orientation
> between residues that are joined/sequenced together is specified - based
> on the prep files.
>
> For example, if I used the command in leap " peptide = seq { ALA ALA } "
> where in the source code would it define how to orient these two residues?
>
> It was clear in the old prep, edit, link style that it was explicitly
> based on overlaying of the dummy atoms with the main chain atoms of the
> previous residue, but I have not explicitly found this in the leap code.
>

The routine you want is ModelAssignTorsionsAround(), in model.c. You will
also want to turn on debug in LEaP (instructions for amber8 here, slightly
different for earlier versions):

cd $AMBERHOME/src/leap/src/leap
make clean
edit the Makefile, to add -DDEBUG to the ".c.o" rule at the top
make t
mv teLeap $AMBERHOME/exe

Then you can run tleap, and debug mode will be turned on. Commands for tleap
would look like this:

debugOn model.c <-- will turn on debug output for routine in that file
loadOff OME.off
loadOff GB.off
b = sequence { OME GB }

At this point, you will see a lot of detail about how the superposition is
being done. I'm not sure whether this will be enough or not. You can also
look at the comments in ModelAssignTorsionsAround(). Be sure that the
hybridization of all of your atoms is defined correctly.

It would be great if you could post a simple case that fails. This should
include the exact leap commands that were used, any input files, and an
example of both the pdb file leap creates _and_ a "good" pdb file that
has the correct coordinates (presumably constructed by some other program).

..hope this helps...thanks for your help....dac