search

rxn4chemistry / rxnmapper

RXNMapper: Unsupervised attention-guided atom-mapping. Code complementing our Science Advances publication on "Extraction of organic chemistry grammar from unsupervised learning of chemical reactions" (https://advances.sciencemag.org/content/7/15/eabe4166).
http://rxnmapper.ai
MIT License
286 stars 68 forks source link

Dubious mapping case #48

Closed parit closed 11 months ago

parit commented 11 months ago

Hello,

I am mapping a reaction smiles as follows:

smiles = "[O-][Cl]=O>>[Cl-].O=O"; RXNMapper().get_attention_guided_atom_maps([smiles])

and the mapped smiles are

[O-:1][Cl+:3][O-:2]>>[O:1]=[O:2].[Cl-:3]

why is the =O in the substrate broken in the mapped smiles?

https://www.rhea-db.org/rhea/21406

parit commented 11 months ago

same problem here:

smiles: "[O-]Cl=O.[Na+].[Na+]>>[O-]Cl=O.[Na+].[Na+]" mapped: "[Na+:1].[Na+:2].[O-:3]Cl+2:4[O-:6]>>[Na+:1].[Na+:2].[O-:3]Cl+2:4[O-:6]"

https://www.rhea-db.org/rhea/71212

avaucher commented 11 months ago

This is likely because the reaction is canonicalized by default. See here for an example with no canonicalization