Enables separate control of molecules separation and random rotations

[amjjbonvin]

You are about to submit a new Pull Request. Before continuing make sure you read the contributing guidelines.

Checklist

• Tests added for the new code
• Documentation added for the code changes
• Modifications / enhancements are reflected on the haddock3 user-manual
• CHANGELOG.md is updated to incorporate new changes
• Does not break licensing
• Does not add any dependencies, if it does please add a thorough explanation

Summary of the Pull Request

Added separate parameters to rigidbody and flexref that control the separation of molecules (separate) and random rotations around their respective center of masses (randrot). These replace the old randorien parameter.

Default for both are true in rigidbody and false in flexref.

Also added a new example to illustrate it: docking-protein-ligand-centered

Related Issue

Closes #1491

[amjjbonvin]

Note that in terms of test, only the validity of the parameters in the new example config files are tested.

[amjjbonvin]

Not in flexref otherwise rigidbody would have no effect! Only in a scenario where the docking happens during the flexible refinement, i.e. without any rigidbody step

[AnnaKravchenko]

Not in flexref otherwise rigidbody would have no effect! Only in a scenario where the docking happens during the flexible refinement, i.e. without any rigidbody step

Yes, already relaized that I was looking in the wrong .md

[AnnaKravchenko]

I runned docking-protein-ligand-separated-test.cfg and docking-protein-ligand-centered-test.cfg to see the difference between the results. Then openned 5 rigid-body models for each run in pymol.

Based on what I see (screenshot attached) the separated vs centered models are backward? I.e. separated models look to me like a pure rotatinon around ligand’s COM, while centered ones are actually have been moved at least a little from the original position?

But when I look at the input files, the ligand is not overlapping with the receptor to begin with, meaning they are being brought together by airs. So I probably misunderstood some part of the process.

[amjjbonvin]

In both cases ligand is overlapping with the protein, which indicates that separate = true did not translate molecules far enoung for the “separated” case?

Both case are docking the ligands. So they should not be separated at the end

But when I look at the input files, the ligand is not overlapping with the receptor to begin with. So I probably misunderstood some part of the process.

Indeed - something is wrong with that receptor. Will update.
Note that this was implemented based on Han's request and he tested it quite extensively.
Just need to update the example. At the end for this example the results are pretty much the same.
But the option should be useful in different scenarios

[AnnaKravchenko]

Can you check is there’s a typo (that is potentially a bug) in src/haddock/modules/sampling/rigidbody/cns/separate.cns , line 40
evaluatee ($zc = $result)?

Same in flexref

[amjjbonvin]

Can you check is there’s a typo (that is potentially a bug) in src/haddock/modules/sampling/rigidbody/cns/separate.cns , line 40 evaluatee ($zc = $result)?

Same in flexref

Typo indeed - will correct - but without consequences as CNS only interprets the first four characters of a command (not variable) :-)

[amjjbonvin]

Note that this typo is also in the main branch

[amjjbonvin]

The last changes to random_rotations come from actually visualising the rotated molecules prior to docking, which made me realise there was another issue.

[AnnaKravchenko]

Combo "separate=true, randrot=false” rotates ligand by 180-ish (along its longest axis?), but I guess it’s a feature, not a bug

[AnnaKravchenko]

Looking good!