[Dock-fans] orientations and conformations

[Scott Brozell]

Hi,

On Mon, Jun 22, 2009 at 04:17:32AM -0700, Ravinder Abrol wrote:
> I am running Dock 6.2 using a flexible ligand for generating many poses to use 
> with our own scoring functions and I want to control the number of 
> conformations that are generated at the end. I am only using bumps and 
> clash_overlap to eliminate poses. I understand the anchor and grow approach 
> but how I can control it with the available input options available is not 
> clear to me. I am providing below results of two runs along with partial 
> input used to generate them. If you can help me make sense of the numbers, 
> that will greatly help.
> 
> >Run1
> Input1:
> max_orientations  1000
> pruning_use_clustering  yes
> pruning_max_orients  100
> pruning_clustering_cutoff  1000
> Output1:
>  Anchors: 1
>  Orientations:  492
>  Conformations:  975
> 
> >Run2
> Input2:
> max_orientations  1000
> pruning_use_clustering  yes
> pruning_max_orients  1000
> pruning_clustering_cutoff  1000
> Output2:
>  Anchors:  1
>  Orientations:  492
>  Conformations:  1023
> 
> Does 492 mean that it found 492 possible placements of the anchor?

Yes, where a placement means
rigidly oriented in the active site (orientation) by matching its
heavy atoms centers to the receptor sphere centers
http://dock.compbio.ucsf.edu/DOCK_6/dock6_manual.htm#AnchorandGrow
i.e., the anchor was docked 492 ways.

> If yes, then how is pruning_max_orients and pruning_clustering_cutoff used?
> I don't understand that in Run2 I specified 1000 for 
> pruning_clustering_cutoff, but at the end I got 1023 conformations.

The above anchor orientations are then ranked according to their score,
spatially clustered by heavy atom root mean squared deviation (RMSD), and pruned
http://dock.compbio.ucsf.edu/DOCK_6/dock6_manual.htm#PruningtheConformationSearchTree
Note that the definitions of pruning_max_orients and pruning_clustering_cutoff
are swapped in the manual; they should be

    (if pruning_use_clustering = yes)

    * pruning_max_orients [100] (int):
      #The maximum number of anchor orientations carried forward in the anchor & grow
      #search (previously named num_anchor_orients_for_growth)
    * pruning_clustering_cutoff [100] (int):
      #The pruning value cutoff for anchor orientations promoted to the
      #conformational search (previously named number_confs_for_next_growth)
      #Note that this is the N_c in the Pruning the Conformation Search Tree
      #equation.

Thus in run2 pruning_max_orients is greater and more Conformations are 
ultimately obtained after growth.

> In other words, is there a way to control the final number of conformations 
> that are generated?

Yes for normal DOCKing calculations; see e.g., num_scored_conformers
http://dock.compbio.ucsf.edu/DOCK_6/dock6_manual.htm#LigandFileOutput
The controls are not designed for preempted running that merely
generates structures.

Scott