Rule not applied when I think it should be.

[willfinnigan]

Hi,

Nice project.

I've got quite a simple rule I'd like to apply which has no chirality specified.

[#6:4][#8:6][#6:2](=[O:3])[#6:1] >> [#6:4][#6:2](=[O:3])[#6:1]

This works fine in most cases, but not for this ring:

C[C@@H]1CCOC(=O)CC1

However, rings one carbon bigger or one carbon smaller work fine.

Is it something to do with it being a symmetrical product? The standard rdkit method works ok.

Thanks for any help!!

Best,
Will

[willfinnigan]

from rdchiral.main import rdchiralReaction, rdchiralReactants, rdchiralRun

rxn = rdchiralReaction('[#6:4][#8:6][#6:2](=[O:3])[#6:1]>>[#6:4][#6:2](=[O:3])[#6:1]')
reactants = rdchiralReactants('C[C@@H]1CCOC(=O)CC1')
outcomes = rdchiralRun(rxn, reactants)

print(outcomes)

[thomasstruble]

You are correct, the outcomes from rdkit has stereochemistry when the symmetry of the product makes the molecule achiral. The lines in the code that deals with this are:

rdchiral/rdchiral/main.py

Lines 498 to 511 in 3b9ae0e

	# Now, check to see if we have destroyed chirality
	# this occurs when chirality was not actually possible (e.g., due to
	# symmetry) but we had assigned a tetrahedral center originating
	# from the reactants.
	# ex: SMILES C(=O)1C[C@H](Cl)CCC1
	# SMARTS [C:1]-[C;H0;D3;+0:2](-[C:3])=[O;H0;D1;+0]>>[C:1]-[CH2;D2;+0:2]-[C:3]
	skip_outcome = False
	if len(tetra_copied_from_reactants) > 0:
	Chem.AssignStereochemistry(outcome, cleanIt=True, force=True)
	for a in tetra_copied_from_reactants:
	if a.GetChiralTag() == ChiralType.CHI_UNSPECIFIED:
	if PLEVEL >= 2: print('Auxiliary reactant atom was chiral, now is broken -> skip outcome')
	skip_outcome = True
	break

[connorcoley]

Hi Will -- Thanks for creating this issue. Viewed in the context of retrosynthesis, this looks like the desired behavior. We wouldn't be able to synthesize C[C@@H]1CCOC(=O)CC1 through that ring expansion; we could only make a racemic mixture since we are starting from the achiral precursor.

If you're interested in changing this behavior, you could consider modifying the lines @thomasstruble referred you to. However, we always handle ambiguity in a manner that makes sense for retrosynthesis throughout the code, so I'd caution you against making just that single change

[willfinnigan]

Ah that makes sense! (sorry, biologist here)

However, enzymatically this appears to be possible.
https://onlinelibrary.wiley.com/doi/epdf/10.1002/cctc.201300604

I'll have a play with the change @thomasstruble suggested.

Thanks both for getting back. Hopefully see you in Bristol next week.