Increasing Discoverability and Connectivity of Scientific Media through Annotation with iCLiKVAL

Scientific media comes in a variety of languages and formats, including journal articles, books, images, videos, blog and database entries, etc. In the case of textual media, there is often additional information such as tables, figures and supplementary data, associated with or embedded in the text. While are there are many good resources for browsing, searching and annotating some of this media, there is not one place where you can search everything, and generalized search engines such as Google do not allow for the type of comprehensive and precise searches that researchers require. To address this, we created iCLiKVAL (http://iclikval.riken.jp/), an easy-to-use web-based tool that uses the power of crowdsourcing to accumulate annotation information for all scientific media found online (and potentially offline). Annotations in the form of key-relationship-value tuples (in any language), added by users with a variety of options, make information easier to find and allow for much richer data searches. Users can create or join common interest groups, both public and private, to annotate related media together as a community. Users can also create and edit their own controlled vocabulary lists, or import established vocabularies such as MeSH and GO. Within the user groups, vocabulary and bookmark lists can easily be shared. iCLiKVAL is an open-access online tool. While the software itself is not open-source, the database is completely searchable, and all of the collected data is freely available to the research community via our API.