JLigand (1.0.37) did not have a problem reading the smiles and displaying the 
structure although the saved coordinates had the quinolone part mangled.

From: Mailing list for users of COOT Crystallographic Software 
[mailto:[email protected]] On Behalf Of Ethan A Merritt
Sent: Tuesday, March 24, 2015 3:26 PM
To: [email protected]
Subject: Coot ligand builder bug interpreting SMILES string


Seems I am having a bad week with coot.



The smiles string

NC1=C2C(N(CC6(COC6)CO)N=C2C3=CC=C(N=C(OC5CC5)C=C4)C4=C3)=NC=N1

describes the compound in the attached file 1671.png.



However when I feed this string to coot it instead constructs the

molecule shown in the file coot-molecule.png.

Note that the terminal hydroxyl has jumped two carbons over

and a CH2 group has been lost altogether.



If I sketch this molecule ab initio in Calculate->Ligand Builder

then when I select "Tidy up" it is mangled into the same

incorrect isomer as the mis-interpreted SMILES string.



Same thing if I ask coot to export a SMILES string and

read it back in again.



Ethan







--

Ethan A Merritt

Biomolecular Structure Center, K-428 Health Sciences Bldg

MS 357742, University of Washington, Seattle 98195-7742


Reply via email to