Semantator: Semantic annotator for converting biomedical text to linked data

Cui Tao, Dezhao Song, Deepak Sharma, Christopher G. Chute
2013 Journal of Biomedical Informatics  
More than 80% of biomedical data is embedded in plain text. The unstructured nature of these text-based documents makes it challenging to easily browse and query the data of interest in them. One approach to facilitate browsing and querying biomedical text is to convert the plain text to a linked web of data, i.e., converting data originally in free text to structured formats with defined meta-level semantics. In this paper, we introduce Semantator (Semantic Annotator), a semantic-web-based
more » ... ronment for annotating data of interest in biomedical documents, browsing and querying the annotated data, and interactively refining annotation results if needed. Through Semantator, information of interest can be either annotated manually or semi-automatically using plug-in information extraction tools. The annotated results will be stored in RDF and can be queried using the SPARQL query language. In addition, semantic reasoners can be directly applied to the annotated data for consistency checking and knowledge inference. Semantator has been released online and was used by the biomedical ontology community who provided positive feedbacks. Our evaluation results indicated that (1) Semantator can perform the annotation functionalities as designed; (2) Semantator can be adopted in real applications in clinical and transactional research; and (3) the annotated results using Semantator can be easily used in Semanticweb-based reasoning tools for further inference.
doi:10.1016/j.jbi.2013.07.003 pmid:23867104 pmcid:PMC4837761 fatcat:tg6z7opkgndlln4bhjetjppoqm