I’ve been blogging about scanning structures, chemical names and quality. Yesterday I blogged about converting an IMAGE of Palytoxin to a connection table. Last night I was playing with the present commercial version of CLiDE. Now I am testing the latest version. All I can say is I am awestruck by my first results.

To see the complexity of Palytoxin check Chemrefer with this search. One of those hits will take you to this paper..check page 2 for Palytoxin. By the way…you just saw the value of Open Access literature and ChemRefer!

Ok, so here’s the question…how long would it take you to draw that (liquidcarbon did such a test already!) . Next question…would you get it RIGHT?

How’s this for a result. Open the Image, convert to connection table with CLiDE, save as molfile, open ACD/ChemSketch and search PubChem to find the structure of Palytoxin. It took about 40 seconds. It can be reduced easily with a few software enhancements…I think that the structure could be selected, converted to a connection table and looked up in PubChem or ChemSpider is less than 15 seconds. the structure is complex…


What’s the bragging rights here…the complete, with all stereo-bonds, exists in PubChem with the same structure, exists in the literature with the same structure, has a molecular formula of C129H223N3O54 was converted WITHOUT an error by Optical Structure Recognition. Now, am I biased? You betcha. I am now working with the development team for this software to provide feedback, tweak workflows and make it better. Based on this experiment I am extremely impressed. This type of performance will stop fragments from being dropped! And, in my world, that’s important!

Stumble it!

2 Responses to “I’m Hard to Impress….I’m IMPRESSED!!! CLiDE can convert Palytoxin without error”

  1. Gary Martin says:

    OK, so the website has nothing to do with science, but rather reflects one of my other passions. No apologies for that.

    Having drawn palytoxin by hand for purposes of a review article some time ago, getting it done in seconds, vs. probably the 20 or so min it takes you to draw the molecule by hand is a Godsend! One of the problems with any big structure is that you never get right how the damned thing is laid out the first time and you spend a lot of time correcting that, dragging things around to make room for something, not to mention the time spent going in and getting all of the stereochemical orientations correct, and then putting in all the OH groups, ether oxygens, and all the rest of it. Then there’s the rather laborious problem of checking the structure to make sure that you got all the stereochemistry right!

    Suffice it to say that I too am impressed!!!

  2. David Bradley says:

    Gary’s exactly right. It takes me ten minutes to get even a simple functionalised benzene ring drawn properly. An automated way of pulling out proper fully-fledged structures from flat documents has to be one of the most important innovations for bringing the old chemical literature kicking and screaming into the second decade of the 21st century. Oh, incidentally, I reviewed an earlier version of Clide for Elemental Discoveries in April 2003 you can find it on my Sciencebase site.


Leave a Reply