Hi,
I typically use HADDOCK for antibody-antigen docking, but I’m currently attempting to dock a receptor protein to a homotrimer. I adapted the methods I use for antibody-antigen docking and renumbered the residues of the homotrimer to create a continuous numbering and renamed all chains of homotrimer to ‘A’. The receptor chain is B. I also utilized an unambig.tbl file to maintain the integrity of the trimer chains and generated an ambiguous restraint file based on the interacting residues.
However, I’ve noticed that the scores in the caprieval file are all positive. Could you provide any insights on why this might be occurring? Additionally, do you have any suggestions for improving the docking results?
Thank you!
Hi, the recipe you’re using sounds correct…did you have a look at the resulting models? do they make sense? is the trimer kept together correctly?
positive scores are not impossible, especially if you provided lots of ambiguous restraints: many of them will not be satisfied, thus giving rise to a high penalty that leads to positive scores.
Another possibility is that the unambiguous restraints are using incorrect distances, therefore leading to high penalties: you may want to check the unambiguous restraint file.
Hope it helps, cheers!
Marco
1 Like
Hi @marco.giulini ,
I saw the resulting models. Some of them meet my criteria and all the chains were together irrespective of using unambig files (I tried both with and without unambig, but kept ambig). The ambig files had residues which did not form a contiguous surface, so maybe that is the problem?
I checked the distances in unambig file for a few residues. Mostly correct.
One result I got looks like this (with the red surface denoting residues put in ambig, receptor in cyan)
Thanks for your insights. It helped a lot 
indeed the “problem” seems to be in the number of active residues defined on the trimer…if you think they all make sense (hard for me to judge from the picture) you can leave them there, the fact that you get positive scores is understandable in this case 
a recommended way to filter out some active residues is to consider only the surface exposed ones, for example running:
haddock3-restraints calc_accessibility –cutoff 0.4 trimer_structure.pdb
1 Like
Thank you for your prompt reply, @marco.giulini.
My intention was to dock the receptor’s strand region to any of the three red regions and maximise interactions between the disordered region and the red parts also while choosing the complex structure. However, it seems that this method does not allow for a “choose any one” option. 
It appears that taking one region at a time will be a better approach. I will also implement your suggested command to retrieve surface residues. Thanks again for your help! 
1 Like
