Mining Biomedical Publications With The LAPPS Grid

Mining Biomedical Publications With The LAPPS Grid Nancy Ide*, Keith Suderman*, and Jin-Dong Kim** *Department of Computer Science, Vassar College Poughkeepsie, New York USA {ide, suderman}@cs.vassar.edu *Database Center for Life Science (DBCLS), Univ. of Tokyo Kashiwa-no-ha Wakashiba, Kashiwa-shi, Chiba, Japan jdkim@dbcls.rois.ac.jp Abstract It is widely recognized that the ability to exploit Natural Language Processing (NLP) text mining strategies has the potential to increase productivity and innovation in the sciences by orders of magnitude, by enabling scientists to pull information from research articles in scientific disciplines such as genomics and biomedicine. The Language Applications (LAPPS) Grid is an infrastructure for rapid development of natural language processing applications (NLP) that provides an ideal platform to support mining scientific literature. Its Galaxy interface and the interoperability among tools together provide an intuitive and easy-to-use platform, and users can experiment with and exploit NLP tools and resources without the need to determine which are suited to a particular task, and without the need for significant computer expertise. The LAPPS Grid has collaborated with the developers of PubAnnotation to integrate the services and resources provided by each in order to greatly enhance the user’s ability to annotate scientific publications and share the results. This poster/demo shows how the LAPPS Grid can facilitate mining scientific publications, including identification and extraction of relevant entities, relations, and events; iterative manual correction and evaluation of automatically-produced annotations, and customization of supporting resources to accommodate specific domains. Keywords: biomedical text mining, LAPPS Grid, PubAnnotation, Open Advancement evaluation, reproducibility 1. Introduction Keeping up with the ever-expanding flow of data and pub- lications is untenable and poses a fundamental bottleneck to scientific progress. The global research community gen- erates approximately 2.5 million new scholarly papers per year (in English only); a new research paper is published every 12 seconds. Current search technologies typically find many relevant documents, but they do not extract and organize the information content of these documents or suggest new scientific hypotheses based on this organized content. It is widely recognized that the ability to exploit Natural Language Processing (NLP) text mining strategies has the potential to increase productivity and innovation in the sci- ences by orders of magnitude, by enabling scientists to pull information from research articles in scientific disciplines such as genomics and biomedicine. The application of NLP techniques can also lead to hypotheses and discov- eries for which there is “hidden” (not explicitly stated) evi- dence in the research literature and enable linking extracted information to form new facts or new hypotheses to be ex- plored further. These methods enable scientists to rapidly identify publications relevant to their own research as well as make scientific discoveries by scouring hundreds of re- search papers for associations and connections (such as be- tween drugs and side effects, or genes and disease path- ways) that humans reading each paper individually might not notice. Up to now, the use of NLP technologies has required con- siderable skill in the field. However, recent development of environments for constructing customizable NLP applica- …

Content truncated to 500 words