Abstract
To facilitate access to relevant text of literature related to data in GlyCosmos, we have developed a collection of annotated literature resources using the agile annotation method supported by the PubAnnotation system. As a proof of concept, we compiled two dictionaries for glycan motifs and epitopes, plus six additional dictionaries for relevant biological entities, covering organisms, phenotypes, diseases, and anatomical locations. Next, we collected all the PubMed abstracts from 15 selected journals, and annotated them based on these eight dictionaries. This resulted in 279,368 annotation instances made to 15,463 abstracts, meaning that we were able to automatically pull glycan motif and epitope annotations related to diseases, taxonomy, etc. from over 15,000 abstracts. All the annotations were converted into Resource Description Framework (RDF) statements to support flexible querying. For users who are not familiar with RDF, we also developed a Web interface in GlyCosmos to visualize the location of the text in publications as well as query templates to personalize queries for specific terms. Pilot searches and analyses suggest that these resources are useful for navigation of relevant contexts of biomedical associations relevant to glycobiology.