Unable to run haddock from docked Cluspro complex

Hi, I am completely new to this. I have submitted a docked complex obtained from Cluspro to Haddock but the run failed. I noticed they mentioned “structure contains gaps” but I have not edited the protein structures so it should be from the original pdb files. Can you kindly suggest how should I fix this issue?

Thank you in advance!

Dear Ashley,

Looking at your log file:

  • Structure contains gap, is just a warning, and you should not worry about it !
  • The issue comes from the last line, that is quite explicit: No contacts found for selection. This means the two chains are not in contact, therefore binding affinity prediction with Prodigy cannot be computed, as there is no binding between the two partners.

Have you looked at your submitted complex on a molecular visualizer (e.g:PyMOL) ?

Cheers

Dear Victor,

Thank you for replying! I have checked the docked complex on Pymol and their energy score is -1091 so technically they should be interacting. I have included the Pymol view below.

But are the chain names correct? You specified A and B

Hi,

Yes, I made sure I categorized the respective proteins into Chain A and B by using the type alter (sele), chain=’A/B’ command in Pymol.

Could you send me your pdb file so I can investigate ?

Sure, how should I send it to you? Maybe you can share your email as I cannot send attachments as a new user.

Hi all,

I attempted once more, this time I removed the chains that were not interacting based on what I saw on Pymol and it went through. I’m wondering if this is the right way? Was this error due to the complex being too large?

Good to know you managed !

I would rather go for a miss selection of the chains to be analyzed by Prodigy.
It is weird.
Here is my email (v.g.p.reys@uu.nl) if you want to send me the full original structure for us to know where the problem comes from

Hi Victor,

Just hopping back in to see if you have managed to take a look at my complex?

Thank you.

Dear Ashley,
Sorry for the late reply.

From the files you sent me, I have seen that the output of ClusPro contains:

  • A receptor: composed of chains A, C, D, F
  • A ligand: composed of chain A

Therefore:

  • Because there is not chain B in your input pdb file, when prodigy is used selecting chains A and B (--selection A B), no contact can be found!
  • Also, having two chain A (one for the ligand and one that is part of the receptor) can only create troubles downstream.

Modifying the receptor to chain A and the ligand to chain B was probably the smartest move to be able to run prodigy.

Maybe by default ligands are always written with chain A out of the ClusPro webserver, therefore I would advise you to rename receptor chains to B, before the submission to the docking next time.

Hi Victor,

Thank you for your response! I really appreciate it.

However, I do have a question on this. Before submitting my pdb files to prodigy, I defined my receptor and ligands as chain A and B, respectively. I have done this using the type alter (sele), chain=’A/B’ command in Pymol. However, I still received the same error message.

Have I approached your suggestions from a wrong direction? Thanks!

The command alter (sele), chain="A/B" in PyMOL (v.2.5.7) is modifying your current selection sele to chain A/B, which is a weird chain name.

Therefore prodigy will complain, as it will not find the chain B in your file.

Instead, you should:

  • select all the chains of your receptor: to create a selection under (sele)
  • alter (sele), chain="B": to modify the chains to B
  • alter (sele), segi="B": to modify the segment ids to B (just to be sure)
  • save the file

I guess this should work.

But having both the ligand and the receptor containing chains A can only create issues.

Remember that you can specify the chains to be analysed in the prodigy server using the Interactor1 and Interactor2 fields in the form.

Having the same chainID in both ligand and receptor is also asking for troubles…

Oh sorry, I wasn’t clear in my previous reply. I used alter sele, chain=‘A’ for receptor and ‘B’ for ligand. I also double checked the proximity and if the chains are identified (as seen in Pymol) but I kept having “no contact found” error for all the complexes I’ve tried. I am wondering would the structural gaps contribute to this error?


The chains defined are shown at the top left and the red in the structures shows the close proximity.

Definitively an issue with your PDB…

Gaps don’t matter.

Why don’t you try our pdb_tools to change the chainID/segIDs?