Open olegursu opened 10 years ago
Hi Oleg, this is a bug. The first one is the one that should be returned. I've just pushed upstream d9a1672 the fix. Please also note that we've also decided to turn off keto-enol by default. To turn it on, please use the following
echo "C=CNCCC1=CC=CC=C1.Cl" | java -Dketo-enol=true -jar dist/lychi-all-v0.1.jar
Hi Trung,
When running structure A (amlodipine) through Lychi with keto-enol option on: echo "CCOC(=O)C1=C(COCCN)NC(C)=C(C1C1=C(Cl)C=CC=C1)C(=O)OC" | java -Dketo-enol=true -jar dist/lychi-all-v0.1.jar
I get structure (B) in smiles which can be a hypothetical tautomer, probably not real
If I run both structures (A) and (B) I get the same hash key, which is not what I want, I guess what I am asking is it possible to have to separate levels for tautomerism handling (1) simple between neighbors atoms like keto-enol, amine-imine, lactam-lactim, etc. and (2) where long range is handled and leave the user the option to select either or both.
Ah yes, therein lies the danger of blindly allowing (long range) keto-enol. Perhaps the thing to do here, as in the case of salt/solvent removal, is to establish a list of "known" keto-enol tautomers and only allow tautomerism if it's known?
I favor a solution based on rules, what format do you prefer SMIRKS, RXN, etc?
Sure, SMIRKS would be great, thanks!
Hi all,
I have a question regarding the output, for example:
produces 2 hash keys:
CC=NCCC1=CC=CC=C1 null N4765661Y-YGFUN3DUM8-Y8HGNSQ89JY-Y8Y5C3KXJ1U2 C\C=N\CCC1=CC=CC=C1 null N4765661Y-YGFUN3DUM8-Y8HGNSQ89JY-Y8Y6J8LJGS98
which one is suppose to be the canonical one?