ADL: Re-Docking

Mari Chikvaidze chikvaidze at gmail.com
Thu Jan 24 18:21:37 PST 2008


Hi Kar and everybody,

It seems that the re-docking experiment, you've been playing with is very
similar to mine. I came across similar difficulties recently, maybe you, or
someone else can help.

For my re-docking, I "took" the ligand out from the PDB file of
protein-ligand complex, paste the ligand coordinates into PRODRG server and
generate PDBQ format of my ligand.(Did you do the same thing, or you draw
the ligand yourself? Did you energy minimize using PRODRG before generating
formated file?) Then I measured the length of my ligand and set the grid box
centered on the macromolecule,
so that dimensions of the box are in correspondence with ligand length, and
the ligand is able to rotate freely within the box.

Centering the grid box around the residue, having hydrogen bond with ligand,
is something I never tried. To me, it sounds strange why the hydrogen bond
was missing in resulting conformation?!

Regarding you questions, the way I checked the correctness of my re-docking
was low RMSD with reference to input ligand structure. In addition, I was
able to reconstruct the interactions between ligand and protein, that was in
a good egreement with interactions in an original protein-ligand complex.

My initial aim was to use docking protocol established by successful
re-docking in order to dock the same ligand into the same protein, but in so
called apo form. The apo (uncomplexed) structure was also available from
PDB. Using the same grid box size as  for re-docking, with same docking
parameters, I was not able to dock the ligand back :( My feeling is that the
ligand I am using is wrong, as it is the one I get from the protein holo
form.

Does anybody have any suggestions?
Thanks in advance,


On Jan 2, 2008 4:50 PM, Sim Kar Seng <karseng76 at yahoo.com.sg> wrote:

> Dear all,
>
> I'm a new Autodock user and is starting on some docking test.I have some
> questions regarding the results of my docking and will like to seek some
> advices.
>
> I've taken the Lipoxygenase with protocatechuic acid complex (PDB: 1n8q)
> for
> a docking test and have tried docking with different grid size and search
> parameter. I've removed all the water molecule and the ligand for the
> macromolecule and use PRODRG to obtain the structure for the
> protocatechuic
> acid (DHB) ligand.
>
> In the original structure, the ligand (DHB) was found to have a hydrogen
> bond to the HIS523 residue. This was found by using the Build command in
> the
> Hydrogen Bonds menu in ADT. One question here is whether this is the
> correct
> method to find out where the ligand is bonding to?
>
> In a few preliminary dockings where I have a larger grid size (as large as
> 126x126x126), none of the docked pose were near to the ligand position in
> the original complex even with higher search parameters like max number of
> evaluations of 25000000, max number of generations of 100000 and ga_run of
> 12. In each case, the results have poor clustering and the ligand were
> away
> from the original location in the complex. However, the best docked energy
> in some of these clusters are around -7.XX.
>
> Subsequently, I've try to restrict the grid map size down to 60x60x60 and
> centered around the the HIS523 residue that the ligand binds to in the
> original complex and perform a docking with the higher search parameters
> as
> above but with a ga_run of 20. The results did converge to a good
> clustering
> and docked pose were quite near to the HIS523 residue that the ligand
> docked
> to in the original complex. However, the docked energy for all pose were
> quite near at around -5.XX. In addition when I use the build command in
> the
> Hydrogen Bonds menu in ADT to build the bond, the bonds were not to the
> HIS523 residue as in the original complex.
>
> There's a few question to the above observation.
> - Is it abnormal/uncommon to find a lower docked energy when performing
> the
> docking with a smaller grid and with higher search parameters settings?
>
> - In this case, the docking with a good clustering is the one closer to
> that
> in the original complex but have a lower docked energy. So in general,
> should I place more emphasis on getting a good cluster than a lower docked
> energy?
>
> - Should I be concern that in the above final docking with the best
> clustering, the hydrogen bond does not match with that in the original
> complex? As a whole, given the above, can I consider the above to be an
> accurate/successful docking or does it mean that the ligand fails to dock
> correctly?
>
> My most sincere apologies for the very long posting as I'm not sure how to
> put it across without missing too much details. Thank you for your
> patience
> and any advice will greatly be appreciated.
>
> Brgds,
> Kar Seng
>
> ________________________________________________
> --- ADL: AutoDock List  --- http://autodock.scripps.edu/mailing_list ---
>



-- 
Mari


More information about the autodock mailing list