Supporting Biological Pathway Curation Through Text Mining

Abstract

Text mining technology performs automated analysis of large document collections, in order to detect various aspects of information about their structure and meaning. This information can be used to develop systems that make it much easier for researchers to locate information of relevance to their needs in huge volumes of text, compared to standard search mechanisms. With a focus on the challenging task of constructing biological pathway models, which typically involves gathering, interpreting and combining complex information from a large number of publications, we show how text mining applications can provide various levels of support to ease the burden placed on pathway curators. Such support ranges from applications that provide help in searching and exploring the literature for evidence relevant to pathway reactions, to those which are able to make automated suggestions about how to construct and update pathway models.