Chemicalize your molecules is a new, free service from ChemAxon. It uses their “name to structure” parsing and structure-based predictions to identify chemical structures from web pages and other text and provide predicted data related to each structure, which seems like a rather neat trick to me. I used it to add some chemical information to a blog post about the monkshood toxin, acinitine, on my Imaging Storm site earlier this week.

ChemAxon’s Alex Allardyce tells me that, “The service allows you to browse the web and see structures for chemical names (text) identified in the web page. For each chemical structure image generated, you can link through to predicted data from the structure.” They’ve also added new data pages so that for each structure users can click through the structure image to arrive at a configurable data page with potentially more than 200 predicted properties. The site’s developers are currently working on new functionality.

You might be wondering whether this site is simply duplicating a site like PubChem or ChemSpider. Apparently not. Allardyce tells me that, “All properties are calculated, no actual data, so results every time. We are parsing the URLs in real time (parsing out the chemical names and annotating the pages with structures) and serving up the page as you would expect. Any links you follow are also chemicalized so your whole experience is chemicalized.”

You can check out a “chemicalized” version of a recent Reactive Reports item on the chemistry of absinthe here or the chemicalized version of my monkshood post. It captures the various chemical names but also seems to tag the word “check” and show a structure, although I’ve reported that as an error, and the site is in the “alpha” phase after all.

The video explains how to use Chemicalize:

Chris Swain has apparently built a Safari chemicalize plugin to improve logins and such.

5 thoughts on “Chemicalize your molecules

  1. Having perfect name to structure conversion will be impossible as the approach is both a systematic name to structure conversion algorithm (and none of them are perfect and MANY names are not even systematic…just pseudo-systematic with ambiguities) and common name dictionary look-ups, and the dictionaries would need to grow daily as new compounds are discovered and given common/trivial names. The Chemicalize website is an excellent offering to the community. Good for ChemAxon…they have done a very good thing for all of us chemists

  2. Depends on what you want from your web pages, Mark. If you’re a chemist or working with molecular structures this is not in the least bit lame.

  3. Just to let you all know that we have added chemical search to the site.

    The search let’s you draw, upload a structure or name a molecule as input and perform various structure searches (substructure, exact match and similarity) against the structures identified from web pages already ‘chemicalized’ by users and see the resulting structures (with the query highlighted in the result).

    From each result structure you can go on to see a list of web pages where the target structure was included. An elegance is that you can also see other structures on the listed pages and select popular or diverse selection, to understand if the page is chemically relevant before you visit the link. Feedback from test users was very positive especially because they can chemically search ‘a significant proportion’ of Wikipedia.

    Other features added include: Auto-complete, structure images for all auto-complete names and roll over large versions of structure images. The name to structure parser is also improved to increase conversion rate and remove error.

    For more information see this demo video at the site or this presentation of and the underlying technologies used from our recent US UGM.

    Hope this is useful

    Alex Allardyce: ChemAxon

Comments are closed.