Text analytics for life science using the Unstructured Information Management Architecture

Abstract

Biomedical text plays a fundamental role in knowledge discovery in life science, in both basic research (in the field of bioinformatics) and in industry sectors devoted to improving medical practice, drug development, and health care (such as medical informatics, clinical genomics, and other sectors). Several groups in the IBM Research Division are collaborating on the development of a prototype system for text analysis, search, and text-mining methods to support problem solving in life science. The system is called "BioTeKS" ("Biological Text Knowledge Services"), and it integrates research technologies from multiple IBM Research labs. BioTeKS is also the first major application of the UIMA (Unstructured Information Management Architecture) initiative also emerging from IBM Research. BioTeKS is intended to analyze biomedical text such as MEDLINE™ abstracts, medical records, and patents; text is analyzed by automatically identifying terms or names corresponding to key biomedical entities (e.g., "genes," "proteins," "compounds," or "drugs") and concepts or facts related to them. In this paper, we describe the value of text analysis in biomedical research, the development of the BioTeKS system, and applications which demonstrate its functions. The large scale sequencing of the human genome has greatly increased our knowledge of the genetic basis of biological processes and accelerated the pace of research and development aimed at treating disease and enhancing the health and well-being of humans. However, these advances also result in increased complexity in understanding and applying biomedical research and data. There is consensus in the life-science (LS) industry and academic laboratories that managing the complexity of biological data and knowledge requires an integrative, informationbased systems approach, in which computer technology must play an essential role. For a cogent analysis of this situation and the role of computational methods in life science, see References 1-3. Key components of computational technology that are relevant to this effort include analyzing, searching, and mining biomedical text, and correlating the structured data derived from texts with data derived from biomedical experiments, transcribed medical records, and so on. This paper describes an IBM Research project to exploit and develop the textanalytical technology needed for managing, analyzing, and using biomedical text to solve problems in life science. We call the system BioTeKS for "Biological Text Knowledge Services." BioTeKS is also one of the first major systems implemented with the IBM Unstructured Information Management Architecture (UIMA), which is described later in this paper, in other papers in this issue, 4 and elsewhere. 5

Extracting sentiment from healthcare survey data: An evaluation of sentiment analysis tools

“… Health: manage and analyse biomedical text that can help in treating diseases and speed up the drugdiscovery process, in order to enhance the health of humans [2].…”

“…In addition, systems like BioTeKS exist that "support problem-solving in life science by analysing biomedical text" [2]. In general, most researchers focus on creating a system or algorithm that is usually fully customisable depending on the goal of the project.…”

Corpus Refactoring: a Feasibility Study

“…Finally, the curated data from the last step was programatically converted into the WordFreak and embedded-XML formats via an application centered around the Unstructured Information Management Architecture (UIMA) [18-20]. Sentences and the associated annotation data were imported into the UIMA framework where they were placed in a standardized data structure, and then outputted in their refactored form by an output-printer component.…”

Semantic search via XML fragments

“…Research in the document-centric view has focused on marrying traditional IR techniques over text spans and tree retrieval techniques over XML tags [5,2], as well as extending document representation to support stand-off annotations [12] that allow for the crossing of tags. In our work, we build upon these results to demonstrate how the XML Fragments query language can leverage a collection of annotated XML documents to address different query-time semantic needs.…”