Comprehensive data mining of the scientific literature has become an increasing challenge. To address this challenge, Elsevier's Pathway Studio software uses the techniques of natural language processing to systematically extract specific biological information from journal articles and abstracts that is then used to create a very large, structured, and constantly expanding literature knowledgebase. Highly sophisticated visualization tools allow the user to interactively explore the vast number of connections created and stored in the Pathway Studio database. We demonstrate the value of this structured information approach by way of a biomarker use case example and describe a comprehensive collection of biomarkers and biomarker candidates, as reported in the literature. We use four major neuropsychiatric diseases to demonstrate common and unique biomarker elements, demonstrate specific enrichment patterns, and highlight strategies for identifying the most recent and novel reports for potential biomarker discovery. Finally, we introduce an innovative new taxonomy based on brain region identifications, which greatly increases the potential depth and complexity of information retrieval related to, and now accessible for, neuroscience research.
Keywords: Pathway Studio; literature analysis; literature data mining; natural language processing.
© 2016 New York Academy of Sciences.