Hi Egon Wrt 2.below - You are not doing anything wrong! I had forgotten that I put an optional clean-up step into the Smiles parsing code (using the Smiles parser/Smiles creator) before the molecule is passed to the dbst. FixAromaticBondOrders routine. This means that the actual Smiles which gets passed on to generate 13 double bonds (4 x 3 + 1) is:
"O=C1Oc6ccccc6(C(O)C1C5c2ccccc2CC(c3ccc(cc3)c4ccccc4)C5)" ie the pyranone ring is no longer marked as aromatic and the algorithm is happy to kekulise the four aromatic rings remaining. If I pass the original Smiles: "Oc1c(C2CC(Cc3ccccc23)c4ccc(cc4)c5ccccc5)c(=O)oc6ccccc16" I get the same result as you do.. ..and this is actually what should happen - every combination of double bonds which gets tried is rejected because of the 'illegal' pyranone ring and so the original molecule with 1 double bond is returned - the same result as I believe the original algorithm would have given if it hadn't timed out first (but < 1 sec vs > 100 sec). Sorry for the confusion. Incidentally, there would be a performance advantage if the isStructureOK routine was modified to only check the local interconnected ring group rather than the complete structure (even better, as stated before, if it just checked the local double bond layout instead of perceived aromaticity). Btw - fun to try your code (either with the modified algorithm or the original CDK version) on the Smiles for Acridine: "c1c2ccccc2nc3c1cccc3". It usually returns 7 double bonds but, occasionally (1 in 30+ occasions), it only returns 6 as discussed previously. Kevin -----Original Message----- From: Egon Willighagen [mailto:[email protected]] Sent: 16 March 2012 06:25 To: Lawson Kevin GBJH Cc: [email protected] Subject: Re: [Cdk-user] Deduce Bond System Tool >> 2. The present algorithm's aromaticity test is not sufficient to demonstrate >> a molecule has all the double bonds layed out correctly > I have added in that git branch a unit test, but I am only getting one double bond... > Kevin, what am I doing wrong? + @Test + public void testLargeRingSystem() throws Exception { + String smiles = "Oc1c(C2CC(Cc3ccccc23)c4ccc(cc4)c5ccccc5)c(=O)oc6ccccc16"; + SmilesParser smilesParser = new SmilesParser(DefaultChemObjectBuilder.getInstance()); + IMolecule molecule = smilesParser.parseSmiles(smiles); + + DeduceBondSystemTool dbst = new DeduceBondSystemTool(new AllRingsFinder()); + molecule = dbst.fixAromaticBondOrders(molecule); + Assert.assertNotNull(molecule); + + molecule = (IMolecule) AtomContainerManipulator.removeHydrogens(molecule); + Assert.assertEquals(34, molecule.getAtomCount()); + + // we should have 14 double bonds + int doubleBondCount = 0; + for (int i = 0; i < molecule.getBondCount(); i++) { + IBond bond = molecule.getBond(i); + if (bond.getOrder() == Order.DOUBLE) doubleBondCount++; + } + Assert.assertEquals(14, doubleBondCount); + } Egon Syngenta Limited, Registered in England No 2710846 Registered Office : Syngenta Limited, European Regional Centre, Priestley Road, Surrey Research Park, Guildford, Surrey, GU2 7YH, United Kingdom This message may contain confidential information. If you are not the designated recipient, please notify the sender immediately, and delete the original and any copies. Any use of the message by you is prohibited. ------------------------------------------------------------------------------ This SF email is sponsosred by: Try Windows Azure free for 90 days Click Here http://p.sf.net/sfu/sfd2d-msazure _______________________________________________ Cdk-user mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/cdk-user
