Quertle - Find What Matters with Semantic Searching of the Biomedical Literature

• GIF

• GIF

• GIF

• GIF

INTRODUCTION Quertle® is an innovative website for searching and investigating the biomedical literature. Using cutting-edge semantic-driven text analytics and concept extraction, Quertle finds conceptual relationships - not just query terms scattered throughout a document - and provides intuitive, dynamic filtering. These capabilities enable the user to quickly find the relevant documents, instead of being overwhelmed by long lists of results. Quertle's integrated search covers licensed content from NLM (PubMed and TOXLINE), NIH Grants, full-text articles from publishers and Open Access sources, news articles curated for life science relevance, and vetted whitepapers and research reports. Where applicable (e.g., a PubMed Central article also in PubMed), the records are merged to bring the unique content for each source together. Created by life science professionals, for life science professionals, Quertle is a free resource for the betterment of research and patient health.   DESIGN INNOVATIONS Relationship-based Searching: Quertle's ground-breaking search is built on semantic and concept extraction methods, fine-tuned for searching the life science literature. Essentially, Quertle semantically examines each sentence to find the conceptual relationships asserted, such as "asbestos causes mesothelioma". The resulting database of over 250 million assertions enables fast searching and filtering on the fly. To see the value of this approach, consider searching for the relationship between diabetes and cancer. With Quertle's semantic search, you would find "Diabetes is well known to be associated with pancreatic cancer" (Perrin et al, 2007, BMC Med 5:25) but not "Curcumin has been used to treat cancer, diabetes and other pathologies" (Tikoo et al, 2008, Br J Pharmacol 153: 1225), even though both sentences contain "diabetes" and "cancer". Thus, the user does not have to pick through irrelevant hits and finds relevant information much faster. Quertle finds relevant assertions in even complicated sentences so that a high recall is achieved. However, unlike other sites that include full-text, Quertle does not search "gray content" such as References, further ensuring the results have assertions relevant to the user's query. Innovative Key Concepts Filter: From each relationship that matches the user's query, Quertle extracts the key concepts for dynamic filtering. The key concepts include verbs, so the relevant actions, as well as the entities, can be discerned. Using concepts from within the relationships found by the search, Quertle maintains its focus on what is relevant to the user. Novel Ontology Applicaton: Quertle improves searching through the use of extensive, manually-curated ontologies, which include synonyms and aliases for genes, chemicals, cell types, and many more entities. This means a user doesn't have to do a search such as "BSG or TCSF or neurothelin or …"; entering any name for this gene would find all synonyms. Taking this further, Quertle uses ontologies to enable searching for classes of objects using Power Terms™. Power Terms represent a class of related concepts. In some cases, such as searching for "neurotransmitter", finding the term neurotransmitter itself along with actual neurotransmitters is desired. In other cases, the class name is so prevalent, such as "disease", that the user may prefer to find only actual disease names. Thus, Quertle introduced special 'find members, not the name' Power Terms, including $Diseases, $Proteins, $Chemicals, and many more. Over 1300 Power Terms are currently supported, including actions (verbs). When the user searches for a Power Term, the Key Concept filter automatically lists members of the class that were found within relationships relevant to the user's query. E.g., in a search for "$Diseases of aging", diseases stated in the literature to be associated with aging would be listed; a click on one would then immediately limit the results to those about that specific disease. Keywords: To assist users make the transition to relationship-based searching, Quertle does do a Keyword search (separate tab) taking advantage of PubMed's MeSH headings and chemical lists - but with the added benefit of Quertle's ontology and Power Terms.   IMPACT ON USERS Quertle enables users to effectively find the relevant literature, saving considerable time - sometimes hours for a single search. In other cases, users find relevant references on Quertle that they could not find at all on other sites. As a result, Quertle has become a key resource for hundreds of thousands of users in 178 countries.   USABILITY Quertle has focused on ease-of-use, providing powerful capabilities in an intuitive interface. Searching: As the user enters a query, Quertle provides search suggestions based on content from the literature, instead of merely popular searches. Default suggestions, including the use of verbs, educate the user about how to take full advantage of Quertle’s powerful search methodology. With a free MyQuertle account, a user can save a search and get automatic alerts when a new relevant document is added. Results: Quertle presents results in an easy-to-understand format, with the relevant relationships shown and search terms highlighted in context. This way, the user can immediately understand why the document was found and assess its importance to their need. The document abstract can also be displayed directly on the results page, again with search terms highlighted. The user can sort the results by relevance or date. Dynamic filtering directly on the results page enables the user to effectively explore the results. Connectivity: Once a user identifies the relevant results, the user can access the full text article by clicking the PDF icon (for open access full-text indexed by Quertle) or the My Library link connected to their institution's subscriptions. More than 1000 universities and other organizations have connected their holdings to Quertle. Quertle also supports export to bibliographic management software.   ACCESSIBILITY Quertle is freely available to anyone in the world. The Quertle site is tested with most modern PC browsers as well as smartphones and tablets.   EXTENSIBILITY Quertle is architected to enable additional content collections (e.g., patents) and additional ontologies to be added over time.   SUMMARY The combination of leading semantic and concept extraction methods provide a powerful, yet user-friendly search engine. The use of such methods is no longer an academic or hypothetical exercise; Quertle is already a critical resource for users around the world.