CCL: (more information) local search in AutoDock

Just to clear up some misunderstandings from my first post:
 I have been using AutoDock 3.05 and ADT for almost a year now, performing
 GA-LS runs for many different ligands and receptors, always using random start
 coordinates for the ligand.  I know that a lot can go wrong with the docking
 box and ligand coordinates.  What I am trying to do is use just a local search
 to minimize the coordinates of a ligand pose from a GA-LS run.  I've double
 checked the coordinates, and in the DPF it says the ligand center is very
 close to the center of my docking box.  My docking box is the same one I have
 used for the past year for docking, and is the same one that produced the pose
 I want to minimize.
 > From the DPF:
 ligand center - 2.515 64.475 68.117
 Grid box center - 3.524 64.788 64.837
 And I still get the problem I outlined below.  From what I can tell, there is
 no reason for the ligand to end up outside the docking box.
 I'm trying to use a local search (pseudo-sw) to minimize a docked pose of a
 ligand into a receptor.  When I look at the ligand and the receptor in ADT,
 the ligand is in the correct starting position.  The Grid box is large enough
 to contain the search space.  I'm using a non-random start, non-random
 relative dihedral offset, and initial relative dihedrals, which should all
 start the local search from the specified initial coordinates.
 The output ligand poses end up completely outside the docking box, and outside
 the receptor as well, with an RMSD of 85 Angstroms!!!  Can anyone tell me what
 is happening, as I see no reason why the ligand should end up outside the
 docking box!
 Thank you for your help,
 Robert Flight
 Masters Student
 Department of Chemistry
 University of New Brunswick
 Fredericton, NB  Canada  E3B 6E2
 e-mail: robert.flight{at}
 "People think it must be fun to be a super genius, but they don't realize
 how hard it is to put up with all the idiots in the world."
 -Calvin from Calvin & Hobbes