Saturday, October 07, 2006

CSLC Versus eMolecules -- Round One

I've finally updated the UsefulChem Molecules blog (subscribe to with BlogLines here) to show eMolecules as the supplier instead of chmoogle.  I've also added a link to perform searches for sub-structures (there are simply too many in most cases to be included directly in the feed).

Using the updated feed, I performed several comparisons between Chemical Structure Lookup Service (CSLC) and eMolecules, and eMolecules came out the clear winner.  For example, for the SMILES
"C(C(NC(=O)OC(C)(C)C)Cc1ccccc1)(=O)O" (which is UC0005 in the UsefulChem Molecules blog), CSLC returned 17 hits, whereas eMolecules gave 46.  Other SMILES gave similar results.  Sheer number of results, of course, may not be the whole story.


At 10:42 AM, Blogger ChemSpiderman said...

JC, I just searched ChemSpider using the "C(C(NC(=O)OC(C)(C)C)Cc1ccccc1)(=O)O" string for the substructure. At present, with a database of 16.5 million compounds, I returned 170 hits in 1.4 seconds. Just fyi, we will roll out a newly improved substructure search engine probably later this week where we are seeing a 30x improvement in search performance. It doesn't mean the search will be 1/30 of 1.4 seconds... a lot of that your own connectivity and bandwidth.

If you'd like any more comparisons simply use the Advanced Search capabilities of ChemSpider.

At 6:50 PM, Blogger ChemSpiderman said...

I just repeated the search without a constraint of number of hits...there are 3410 structures with that structure on the ChemSpider index. A little more impressive. I'm interested in outlining a set of criteria by which to compare such services if you have time.

We are about to release ChemSpider "Gamma"...moving from beta. You'll like a lot of the new capabilities I believe.

Your problems re. the applet searching should have been solved this afternoo...assuming you were on IE 6.


Post a Comment

<< Home

Creative Commons Attribution Share-Alike 2.5 License