On 22 June 2010 03:31, Craig A. James <cja...@emolecules.com> wrote: > On 6/21/10 6:32 PM, Tim Vandermeersch wrote: >> On Tue, Jun 22, 2010 at 3:17 AM, Geoffrey Hutchison >> <ge...@geoffhutchison.net> wrote: >>>> The changes are in svn trunk. The current test set (aromatics.smi, >>>> attype.00.smi& nci.smi) is not good though. >>> >>> By that, I assume you mean that the current test set is not sufficient for >>> SSSR vs. LSSR? I can probably generate a *MUCH* larger test set if that's >>> what you need. >> >> There are not enough cases where the SSSR is not canonical. I already >> added some (tetrahedron, cube, octahedron, fullerene 20& 60). While >> the SSSR doesn't fail the canonical test for bridged rings, having 3 >> rings in the LSSR (2 in SSSR) gives more intuitive results for >> substructure search. > > An interesting way to test various things is with the SMILES "anti-canonical" > option, -xC. It's amazing how many algorithms can have hidden atom-order > dependencies. Unfortunately, it's hard to make test cases using -xC directly > because the results are deliberately quasi-random. > > Craig
Just a quick note, RDKit and OEChem both have something to say on the subject of SSSR. See for example links and discussion at http://ctr.wikia.com/wiki/Ring_counts_in_a_SMILES_file. ------------------------------------------------------------------------------ ThinkGeek and WIRED's GeekDad team up for the Ultimate GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the lucky parental unit. See the prize list and enter to win: http://p.sf.net/sfu/thinkgeek-promo _______________________________________________ OpenBabel-Devel mailing list OpenBabel-Devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openbabel-devel
