CHEMSPIDER: A PLATFORM FOR CROWDSOURCED COLLABORATION TO CURATE DATA DERIVED FROM PUBLIC COMPOUND DATABASES - Collaborative Computational Technologies for Biomedical Research - page 377

Figure 22.4 ChemSpider is a multimedia container supporting photographic images,
MP3 fi les, and YouTube videos. An embedded video showing fast photochromism is
displayed. The video is associated with the ChemSpider record http://www.chemspider.
com/21230378#description.
directly to a publication. For those articles that cannot be linked using a
Pubmed ID or DOI, a user can simply add the full reference details (see Fig.
22.5) and the data will be viewable directly in the record.
When ChemSpider joined RSC [82], a path was initiated to integrate RSC
content into the database. RSC had previously developed an award-winning
semantic markup project known as Project Prospect [83] . “ Prospected ” articles
incorporate standard metadata within the full text of their articles and combine
this with an intuitive on-screen manifestation of the advantages of including
these metadata. For chemists this translates to a number of features, including
the display of compound pages showing the chemical structure, various identi-
fi ers, and links to other online resources when hovering over a chemical name.
Chemical structures which are prospected in the articles are now deposited
directly into ChemSpider on an ongoing basis together with a direct link back
to the associated article. This makes RSC articles more discoverable and pro-
vides direct benefi ts to the reader of the article as the compounds are linked
into the ChemSpider database, thereby opening up access to an expansive set
of data and links across the Internet.
Literature linking has been established in a more automated fashion taking
advantage of freely available APIs and the ongoing curation work underway
on the database to produce a validated dictionary of chemical names associ-
ated with the chemical records. Validated chemical names are used as the basis
of a search against the Pubmed database searching only against the title and
the abstract. In this way a search on cholesterol, for example, would only