Ensemble vs separate docking runs


#1

The HADDOCK category is meant to discuss any HADDOCK-related issue. For general information about HADDOCK refer to http://www.bonvinlab.org/software/haddock2.2

Hello Everyone

I am trying to dock a protein domain (Protein 2) to a site on another protein containing a flexible loop (Protein 1). I would like to get an idea about the favourable conformations of the loop for protein-protein docking. In order to do so I tried the following.
a) Creating an ensemble of cluster representative conformations(12 representatives) of protein 1 with different loop conformations and then using this ensemble to dock protein 2 on it.
b) Individually docking each conformation of protein 1 with protein 2.

The restraints used were exactly same for all the calculations.

I have following queries regarding results.

  1. Is it possible/correct to compare the results of 12 separate runs considering that the restraints etc were exactly same for all the runs with only difference being conformations of protein 1?
  2. Is it possible to combine the results of these 12 runs and score the complexes again? If yes, how?
  3. Would increasing the conformational sampling in approach (a), provide a better idea about the relative preference of certain conformation as opposed to approach (b)?
  4. *rmsd.disp file is missing in the downloaded results for most of the runs. Is there a way to calculate RMSDs for the complexes to perform clustering using different cutoffs?

Thank you


#2

My answers to your questions:

  1. -> YES it does make sense

  2. For that you would have to download all the runs (if you used the server), combined the PDB files from the water directory (but you will need to rename them for each run to avoid overlap) and then perform a manual analysis. Refer for that to:

    http://www.bonvinlab.org/software/haddock2.2/analysis/#manual

  3. Yes it would make sense, but you won’t be able to generate that many water-refined models. You could push however to say 10000/400/400 for the three stages

  4. It is not missing… it is not there since the clustering default is now FCC. Instead there should be another file called complex_fcc.disp


#3

Thank you Alexandre.