Complicated metal complexes as ligand#273
Complicated metal complexes as ligand#273rwxayheee wants to merge 6 commits intoforlilab:developfrom
Conversation
double, and more bond types; M-L-M structure
rwxayheee
changed the title
| for idx in neigh_idx_to_nr_h: | ||
| n = neigh_idx_to_nr_h[idx] | ||
| newidx = idx - sum([i < idx for i in idx_to_rm]) | ||
| mol.GetAtomWithIdx(newidx).SetNumExplicitHs(n + 1) |
There was a problem hiding this comment.
SetNumExplicitHs without Resetting it (since the added hydrogens are non-real) might be causing some problems when functions like RemoveHs are called later on the ligand mol
There was a problem hiding this comment.
Instead of resetting it, not modifying atoms in input mol is better
|
This PR allows preparation of metal complexes with dative bonds as ligands. An example is heme (with Fe(II): Input SDF: Output PDBQT: |
rwxayheee
changed the title
|
Merged develop into this branch. Added a test (example showed in this PR):
|
meeko/molsetup.py
Outdated
| rm_to_neigh[atom.GetIdx()].add(neigh.GetIdx()) | ||
| if not idx_to_rm: | ||
| return Chem.Mol(mol), idx_to_rm, rm_to_neigh | ||
| return input_mol, idx_to_rm, rm_to_neigh |
There was a problem hiding this comment.
Need to either return a copy of input_mol, or make a copy before changing Se to S for computing Gasteiger charges in the function that calls remove_elements()
There was a problem hiding this comment.
Can you explain why and what's the difference?
There was a problem hiding this comment.
It's to prevent a side effect (a modification) on a molecule that a user passes to an instance of MoleculePreparation. This could be a padded_mol in a Polymer. The modification is changing Selenium to Sulfur here: https://github.com/rwxayheee/Meeko/blob/cb7b4e7db4334f161b35124135214ca69099c3f1/meeko/molsetup.py#L1768-L1771
That's why the original code returns a copy of the input molecule, instead of returning the input molecule itself. Then the copy is modified, and the original input mol is safe from side effects.
There was a problem hiding this comment.
Got it, thanks, I understand now
changes outside of the function in the subsequent steps
2 participants
This small edit on the charge method allows potentially complicated metal complexes (with datives, double, triple or quadruple bonds) to become ligand, where metal gets a partial charge as the sum of the metal's formal charge (from input) and the Gasteiger charges of non-real Hs that were added to compensate breaking M-L bonds.
It should also support charge assignment in bridged structures like M-L-M. Here, the coordinate atom L will have multiple added Hs, and the charge will be distribute to both metal atoms based on the bond order.
Dative bonds, are a special case. Breaking a dative bonds do not result in addition of non-real Hs.
This process looks complicated. I put some more in-line comments, but in future we could rewrite it.