#1252 InChI generation fails for some compounds

cdk-1.4.x
closed
nobody
9
2013-06-02
2012-09-11
No

Reproduce:
Create a new molecule from SMILES using this smile:
COc1ccc2[C@@H]3C@HC(C)(C)OC4=C3C(=O)C(=O)C5=C4OC(C)(C)[C@@H]6COc7cc(OC)ccc7[C@H]56

Run it thru a metod that sets the aromatics and then remove the ISAROMATIC-flag.
Now generate the molecules InChI with help of the InChI-generator.

The InChI I got
InChI=1S/C32H32O8/c1-31(2)19-13-37-21-11-15(35-5)7-9-17(21)23(19)25-27(33)28(34)26-24-18-10-8-16(36-6)12-22(18)38-14-20(24)32(3,4)40-30(26)29(25)39-31/h7-12,19-20,23-24H,13-14H2,1-6H3

Expected InChI:
InChI=1S/C32H32O8/c1-31(2)19-13-37-21-11-15(35-5)7-9-17(21)23(19)25-27(33)28(34)26-24-18-10-8-16(36-6)12-22(18)38-14-20(24)32(3,4)40-30(26)29(25)39-31/h7-12,19-20,23-24H,13-14H2,1-6H3/t19-,20+,23+,24-

Is it something more settings that I need to do before I generate the InChI? Or is it something more I should remove from the molecule?

Discussion

  • Egon Willighagen

    Yes, I am aware of this problem. The reason is that the new stereochemistry model introduced when I was in Uppsala, used by the SMILES parser, is not yet supported by the InChI code...

     
  • John May

    John May - 2013-06-02

    Avoiding sf markup

        COc1ccc2[C@@H]3[C@H](COc2c1)C(C)(C)OC4=C3C(=O)C(=O)C5=C4OC(C)(C)[C@@H]6COc7cc(OC)ccc7[C@H]56
    
     
  • John May

    John May - 2013-06-02
    • status: open --> closed
     
  • John May

    John May - 2013-06-02

    Fixed now, after running through FixBondOrdersTool but I now get:

    InChI=1S/C32H32O8/c1-31(2)19-13-37-21-11-15(35-5)7-9-17(21)23(19)25-27(33)28(34)26-24-18-10-8-16(36-6)12-22(18)38-14-20(24)32(3,4)40-30(26)29(25)39-31/h7-12,19-20,23-24H,13-14H2,1-6H3/t19-,20+,23+,24-