Knowledge acquisition from the biomedical literature

Abstract

This article focuses on knowledge acquisition from the biomedical literature, and on the infrastructure, specifically text mining, needed to access, extract and integrate the information. The biomedical literature is the major repository of biomedical knowledge. It serves as the source for structured information that populates biological databases, via the process of expert distillation (or curation) of the literature. Today, the literature has grown to the point where an individual scientist cannot read all the relevant literature, and curators of the major biological databases have trouble keeping up to date with newly published articles. Furthermore, important biomedical applications, such as drug discovery and analysis of high-throughput data sets, are dependent on integration of all available information from both biological databases and the literature. The article reviews these applications, focusing on the role of text mining in providing semantic indices into the literature, as well as the importance of interactive tools to augment the power of the human expert to extract information from the literature. These tools are critical in supporting expert curation, finding relationships among biological entities, and creating content for a Semantic Web.