Chemical-text hybrid search engines

presentation · 8 years ago
by Yingyao Zhou (Genomics Institute of the Novartis Research Foundation)

Is there an informatics tool that can accurately identify all documents within a corporate network that describe certain associations between a chemical compound (e.g., a Gleevec analog) and a therapeutic application (e.g., chronic myelogenous leukemia (CML))? Existing solutions, based on either text or chemical search engines alone, suffer from the inclusion of “false negative” and “false positive” search results. We implemented a chemical-text hybrid search engine that is capable of answering the question described above with a complex query: “[Gleevec > 0.8] NEAR CML”. (PubMed ID: 20047295)