ADL: Speeding up Rigid Docking with a Surface

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

ADL: Speeding up Rigid Docking with a Surface

Daniel Kozuch
Hello,

I'm a new user to AutoDock Vina, and I am hoping to dock a rigid protein
structure to a rigid crystal surface.

My strategy was to prepare the input files in AutoDock Tools using the
crystal surface as the "protein" and the protein as the "ligand". I
followed the same steps as the AutoDock Vina video tutorial except I set
all torsions to non-rotatable for my protein, and I changed the grid box to
the space above my crystal surface.

When I began the docking in AutoDock Vina, I expected the docking to be
relatively rapid (since everything is rigid) but the docking is taking more
than a couple hours to complete on 8 cores. I suspect it is because I had
to define a large search space to fit my protein into (50, 50, 40).

Does anyone have suggestions for speeding this process up?

Thanks,
Dan
________________________________________________
--- ADL: AutoDock List  --- http://autodock.scripps.edu/mailing_list ---
Reply | Threaded
Open this post in threaded view
|

Re: ADL: Speeding up Rigid Docking with a Surface

Diogo Martins
Hi Dan,

I suspect that the limiting factor is the number of atoms in your very
large ligand. I think that the Vina scoring function is more or less
insensitive to interactions beyond a certain distance, let's say 5
angstroms. One idea would be to remove all ligand atoms more than 5
angstrom away from the protein (ligand) surface, producing a hollow protein
(ligand). I'm not sure if this is sufficiently accurate or if would be
significantly faster, but may be worth trying.

Another thing that comes to mind: if the crystal surface is totally planar,
maybe you can use a small patch of the crystal surface as the ligand, and
the protein as the receptor.

Good luck,
________________________________________________
--- ADL: AutoDock List  --- http://autodock.scripps.edu/mailing_list ---
Reply | Threaded
Open this post in threaded view
|

Re: ADL: Speeding up Rigid Docking with a Surface

Daniel Kozuch
Thanks for the fast response. Unfortunately, I am using a molecular crystal
(actually water ice in this case) and when fed in as a ligand to Vina, only
the first molecule is considered. Is there a way to keep the entire ice
sheet as a ligand?

Thanks,
Dan

On Thu, Sep 14, 2017 at 9:43 PM, Diogo Martins <[hidden email]>
wrote:

> Hi Dan,
>
> I suspect that the limiting factor is the number of atoms in your very
> large ligand. I think that the Vina scoring function is more or less
> insensitive to interactions beyond a certain distance, let's say 5
> angstroms. One idea would be to remove all ligand atoms more than 5
> angstrom away from the protein (ligand) surface, producing a hollow protein
> (ligand). I'm not sure if this is sufficiently accurate or if would be
> significantly faster, but may be worth trying.
>
> Another thing that comes to mind: if the crystal surface is totally planar,
> maybe you can use a small patch of the crystal surface as the ligand, and
> the protein as the receptor.
>
> Good luck,
> ________________________________________________
> --- ADL: AutoDock List  --- http://autodock.scripps.edu/mailing_list ---
>
________________________________________________
--- ADL: AutoDock List  --- http://autodock.scripps.edu/mailing_list ---
Reply | Threaded
Open this post in threaded view
|

Re: ADL: Speeding up Rigid Docking with a Surface

Diogo Martins
All atoms must be inside a single instance of ROOT / ENDROOT keywords. This
should fix it, I think.

On Sep 14, 2017 7:30 PM, "Daniel Kozuch" <[hidden email]> wrote:

Thanks for the fast response. Unfortunately, I am using a molecular crystal
(actually water ice in this case) and when fed in as a ligand to Vina, only
the first molecule is considered. Is there a way to keep the entire ice
sheet as a ligand?

Thanks,
Dan

On Thu, Sep 14, 2017 at 9:43 PM, Diogo Martins <[hidden email]>
wrote:

> Hi Dan,
>
> I suspect that the limiting factor is the number of atoms in your very
> large ligand. I think that the Vina scoring function is more or less
> insensitive to interactions beyond a certain distance, let's say 5
> angstroms. One idea would be to remove all ligand atoms more than 5
> angstrom away from the protein (ligand) surface, producing a hollow
protein
> (ligand). I'm not sure if this is sufficiently accurate or if would be
> significantly faster, but may be worth trying.
>
> Another thing that comes to mind: if the crystal surface is totally
planar,
> maybe you can use a small patch of the crystal surface as the ligand, and
> the protein as the receptor.
>
> Good luck,
> ________________________________________________
> --- ADL: AutoDock List  --- http://autodock.scripps.edu/mailing_list ---
>
________________________________________________
--- ADL: AutoDock List  --- http://autodock.scripps.edu/mailing_list ---
________________________________________________
--- ADL: AutoDock List  --- http://autodock.scripps.edu/mailing_list ---
Reply | Threaded
Open this post in threaded view
|

Re: ADL: Speeding up Rigid Docking with a Surface

Daniel Kozuch
Thanks so much, that did the trick.

Best,
Dan

On Thu, Sep 14, 2017 at 10:55 PM, Diogo Martins <[hidden email]>
wrote:

> All atoms must be inside a single instance of ROOT / ENDROOT keywords. This
> should fix it, I think.
>
> On Sep 14, 2017 7:30 PM, "Daniel Kozuch" <[hidden email]> wrote:
>
> Thanks for the fast response. Unfortunately, I am using a molecular crystal
> (actually water ice in this case) and when fed in as a ligand to Vina, only
> the first molecule is considered. Is there a way to keep the entire ice
> sheet as a ligand?
>
> Thanks,
> Dan
>
> On Thu, Sep 14, 2017 at 9:43 PM, Diogo Martins <[hidden email]>
> wrote:
>
> > Hi Dan,
> >
> > I suspect that the limiting factor is the number of atoms in your very
> > large ligand. I think that the Vina scoring function is more or less
> > insensitive to interactions beyond a certain distance, let's say 5
> > angstroms. One idea would be to remove all ligand atoms more than 5
> > angstrom away from the protein (ligand) surface, producing a hollow
> protein
> > (ligand). I'm not sure if this is sufficiently accurate or if would be
> > significantly faster, but may be worth trying.
> >
> > Another thing that comes to mind: if the crystal surface is totally
> planar,
> > maybe you can use a small patch of the crystal surface as the ligand, and
> > the protein as the receptor.
> >
> > Good luck,
> > ________________________________________________
> > --- ADL: AutoDock List  --- http://autodock.scripps.edu/mailing_list ---
> >
> ________________________________________________
> --- ADL: AutoDock List  --- http://autodock.scripps.edu/mailing_list ---
> ________________________________________________
> --- ADL: AutoDock List  --- http://autodock.scripps.edu/mailing_list ---
>
________________________________________________
--- ADL: AutoDock List  --- http://autodock.scripps.edu/mailing_list ---