ADL: Rotating Grid Box and Sensitivity Tweaks

Andrew Fant fant at
Mon Aug 6 13:18:34 PDT 2007

Good Afternoon All,
   I have two autodock questions that I could appreciate some feedback on, one
relatively simple, and one I fear that may not be.

Is there an easy, graphical method to rotate the orientation of the grid box
that is set up in ADT?  I realize that this might be non-trivial because by this
point the pdbqt files for receptor and ligand have already been set up using
that coordinate frame, but it would be nice if possible.  I have a model of a
receptor with a long rectangular binding region that naturally runs along
essentially the diagonal of a binding cube, and I spend a lot of time doing dock
searches in areas that aren't particularly useful.

    My other question is if there is some way to tweak the parameters of the
scoring function to increase sensitivity towards weakly binding ligands. I
understand that the 2.5 kcal/mol uncertainly may well be intrinsic to the model.
   It appears to me that there is an intrinsic bias towards larger molecules in
calculated binding energies, since there is more opportunity for VanDerWalls
interactions.  What I am finding when I look at smaller ligands, such as
dipeptides and such, is that autodock will find several candidate binding poses
that all sit at about the same energy and that appear to ignore hydrogen bonding
and ionic bonding entirely and just look at surface complimentarity and "grease
glue".  I understand that a poorly binding ligand won't show much in the way of
interactions with a receptor, but is there a way to make whatever reactions
there are clearer?  I may not have put this well, so feel free to write and ask
me to explain myself better if it would help.


More information about the autodock mailing list