Menu

#906 incorrect smiles generated for alpha-pinene-like structure

cdk-1.6.x
closed
5
2016-08-11
2009-04-30
No

The attached molfile is a quite different structure to alpha-pinene, yet the exact same smiles is generated for both. They are structural isomers (and have the same number of hydrogens attached to each carbon), but different connectivities.

The inchi of the non-alpha-pinene is InChI=1/C10H16/c1-6-4-9-5-10(7(6)2)8(9)3/h8-10H,4-5H2,1-3H3

Discussion

  • Rajarshi Guha

    Rajarshi Guha - 2009-04-30

    The mol file seems to be a CML file and when converted to SMILES via OB, Daylight Depict reports errors:

    WARNING: Atom has unusual valence 6 (normal 4) (dy_rmbord)
    WARNING: [12C]1(=12C[1H])... (dy_rmbord)
    WARNING: ^^^^^^^ (dy_rmbord)
    WARNING: Atom has unusual valence 7 (normal 4) (dy_rmbord)
    WARNING: ...[1H])[1H])[1H])(12CH3([1H])[1H])... (dy_rmbord)
    WARNING: ^^^^^^^ (dy_rmbord)
    WARNING: Atom has unusual valence 7 (normal 4) (dy_rmbord)
    WARNING: ...[1H])([1H])[1H])12CH3([1H])[1H])... (dy_rmbord)
    WARNING: ^^^^^^^ (dy_rmbord)
    WARNING: Atom has unusual valence 7 (normal 4) (dy_rmbord)
    WARNING: ...[1H])([1H])[1H])12CH3([1H])[1H]... (dy_rmbord)
    WARNING: ^^^^^^^ (dy_rmbord)

     
  • gilleain maclean torrance

    Well alright, when I said 'molfile' I really /meant/ 'cml file'... :)

    By the way, the pipeline of conversion would have been:

    molgen->cdk->bioclipse->babel->daylight

    but anyway, I suspect that this is babel's fault, for not converting cml to smiles properly. It seems to do alright with the inchi, though.

     
  • gilleain maclean torrance

    actual molfile of fake 'alpha-pinene'

     
  • Rajarshi Guha

    Rajarshi Guha - 2009-04-30

    Sent a patch that employs the mol2 file of the isomeric alpha pinene - works fine (i.e., different SMILES compared to alpha pinene)

     
  • Rajarshi Guha

    Rajarshi Guha - 2009-04-30

    I meant a patch to include a unit test to check for this bug

     
  • Egon Willighagen

    • status: open --> closed
    • Group: --> cdk-1.6.x
     
  • Egon Willighagen

    I assume this is fixed in CDK 1.5.

     
  • John May

    John May - 2016-08-11

    ~~~~~
    https://cdkdepict-openchem.rhcloud.com/depict/bow/svg?smi=C1(=C(C(C2(C(C1(C2([H])[H])[H])(C([H])([H])[H])[H])[H])([H])[H])C([H])([H])[H])C([H])([H])[H]&abbr=on&suppressh=true&showtitle=false&zoom=1.3&annotate=none
    ~~~~

     

    Last edit: John May 2016-08-11