Thursday, November 1, 2007

Open collection of crystal structures

Following the supposition that crystal structures are facts, and therefore not subject to copyright, Nick Day, while a PhD student, developed CrystalEye. CrystalEye collects the Crystallographic Information Framework (CIF) files from published articles. Using a web spider, CrystalEye gathers these files from five publishers: Acta Crystallographica, American Chemical Society, Chemical Society of Japan, Elsevier, and Royal Society of Chemistry. Currently, nearly 100,000 structures reside on CrystalEye. Future plans include aggregation from institutional repositories.

CrystalEye provides several access points to its data. You can search either by using simplified molecular input line entry specification (SMILES) or cell parameters. In addition, you can browse by journal and issue. Results are neatly packaged in a clickable chart form (see image below--click on image to see it larger).


Clicking the summary will yield more information, such as the DOI number of the article containing the structure, data collection parameters, refinement results, as well as a javascript application that allows you to change the number of unit cells along an axis and view the structure.

As a big proponent of using RSS feeds to collect information that personally appeals to you, I was thrilled to see the RSS feed function available. You can subscribe to a feed by journal, compound class, atoms, or bonds. For example, you can subscribe to a feed that will notify you of any CIF files added to CrystalEye that contain a bond between aluminum and silver.

CrystalEye also create histograms for bond lengths involving different atoms. The histograms are clickable so you can be linked to the structures that fit the specific parameters you clicked on.

For people interested in the development, technology, and future plans for the beta site, Day has a wiki page with more information.

No comments: