Implement ensemble support for docking in CNS-based modules

[rvhonorato]

This PR implements ensemble support for docking and closes #148.

Note that for this to work I had to change the data structure of topoaa, meaning that all modules that come after it need to be refactored, meaning meaning that we can no longer do refinement runs (but I'm not sure if this was possible anyway topoaa > flexref), let's work on this example in another PR.

[codecov-commenter]

Codecov Report

Merging #152 (888e220) into main (f33ef44) will decrease coverage by 0.28%.
The diff coverage is 3.61%.

@@            Coverage Diff             @@
##             main     #152      +/-   ##
==========================================
- Coverage   39.14%   38.86%   -0.29%     
==========================================
  Files          39       39              
  Lines        2174     2195      +21     
==========================================
+ Hits          851      853       +2     
- Misses       1323     1342      +19     

Impacted Files	Coverage Δ
src/haddock/modules/refinement/emref/__init__.py	22.97% <0.00%> (+0.30%)	⬆️
src/haddock/modules/refinement/flexref/__init__.py	22.97% <0.00%> (+0.30%)	⬆️
src/haddock/modules/topology/topoaa/__init__.py	22.22% <0.00%> (-0.57%)	⬇️
src/haddock/modules/sampling/rigidbody/__init__.py	18.94% <4.34%> (-3.14%)	⬇️
src/haddock/libs/libontology.py	45.76% <33.33%> (-0.67%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update f33ef44...888e220. Read the comment docs.

[joaomcteixeira]

You did upgrade flexref and emref. Do these work already, or still need further implementations? What about mdref?

[amjjbonvin]

topoaa -> flexref was anyway not working since flexref expects a PDB containing all molecules… Did not check what you changed, but one way around this would be to actually create already in topoaa the starting complexes (as done in haddock2.4) rather than loading the individual PDB files into the sampling stage. But this might not be handy for third party sampling software. An alternative would be to add a new module to create the starting complexes by combination of input PDB files.

Note that for this to work I had to change the data structure of topoaa, meaning that all modules that come after it need to be refactored, meaning meaning that we can no longer do refinement runs (but I'm not sure if this was possible anyway topoaa > flexref), let's work on this example in another PR.

Can you explain in a nutshell the difference in the structure of topoaa?

[amjjbonvin]

mdref is on my todo list - currently broken at probably both the python and CNS levels

[rvhonorato]

Did not check what you changed, but one way around this would be to actually create already in topoaa the starting complexes (as done in haddock2.4) rather than loading the individual PDB files into the sampling stage. But this might not be handy for third party sampling software. An alternative would be to add a new module to create the starting complexes by combination of input PDB files.

An alternative would be to add a new module to create the starting complexes by combination of input PDB files.

Good suggestions, lets look into this when we implement refinement workflows.

---

Can you explain in a nutshell the difference in the structure of topoaa?

Say PDBObject is an internal object that contains the path to the .pdb and to the .psf

Before, topoaa was outputting a list (all internally), for MolA with 1 model and molB with 2 models:

[PDBObject, PDBObject, PDBObject]

And now it outputs a dictionary

{0: [PDBObject], 1: [PDBObject, PDBObject]}

So the modules that are loading from it need to handle the "multiple-objects-per-molecule" logic instead of simply reading a list.

[rvhonorato]

You did upgrade flexref and emref. Do these work already, or still need further implementations? What about mdref?

These work already!

[joaomcteixeira]

I kinda like that dict structure. Seems very versatile for what we need, present and future.