autodock-placing ligand ; mutate structure

Hi CCLers,
 I have a question regarding Autodock usage. Does anybody find any significant
 difference in the docking result when it comes to placing the ligand in the grid
 box? I found that when I place my ligand at the corner end of the box, I can
 find better interactions between the docked ligand and protein as compared to
 placing the ligand at the binding site itself. The energy of binding found was
 lower and there seemed to be more conformations with interactions as found in
 the crystal structure.
 Secondly, I would like to mutant one residue of my protein manually by
 substituting the residue with another. Can anybody suggest how I may optimize
 the whole protein structure or do I need to optimize it at all? I don't expect
 any significant change in the structure of the protein but i'm not quite sure if
 I can just optimize the residue itself without affecting the whole protein. My
 concern is that the side chain of the new residue may not be accurately
 represented without some form of refinement.
 Thank you kindly for any suggestions/feedback.
 mail2web - Check your email from the web at .