Finding the right information in the literature can be tedious. Common search engines retrieve publications based on the mere presence of your search terms somewhere in the text. It is up to you to review the hitlist for the actual information needed.
But what if a search engine recognized the context of your search based on the biologically relevant connection between your search terms? And then retrieved a list of facts that match that connection, plus the source document and sentence. Introducing New Biology Search or NBS.
How does it work?
MedScan® NLP is a proprietary natural language processing technology we use at Elsevier to automatically extract biologically relevant facts from published literature. Using syntactic and semantic analyses, the system identifies and extracts essential facts based on predefined properties and interrelationships of entities in the biomedical literature. So the sentence “We have shown that ETS1 can stimulate GM-CSF in Jurkat T cells” becomes the standardized database structure Activate(ETS1, GM-CSF), and a search for “proteins activated by ETS1” retrieves this and other records containing evidence of a protein being activated by ETS1.
MedScan NLP currently extracts high-quality contextual facts from 28.6M Medline® abstracts and 10,000 journal titles, and has generated a repository of 10.6M unique relations (biological facts) supported by 56.6M references.
What we would like you to do
Watch the video here that shows you how NBS works, and then perform as many searches as you can think of, using our sample queries as a guide. As you explore the tool, consider whether the system is interpreting your search terms as you expect, browse through results and explore the resulting data tables and graphs. In the end, please fill out the short questionnaire here.
Click to expand the list of sample queries under the NBS search box above. Select any search of interest to transfer it to the search box and click Search. Feel free to also substitute terms in the query.
On the resulting page, results are tabulated for an easy overview of the relations found. Each row highlights a hit, the type of relation between the two entities, and the direction of that relation. If known, mechanism underlying the relation is also listed. Click any entity or relation to reveal more detailed information.
Click Open Graph View to reveal a graphical representation of the results. You can also switch between different views using "View" button in the upper right corner.