An Approach towards Semi-automated Biomedical Literature Curation and Enrichment for a Major Biological Database

Abstract

As part of a large-scale biocuration project, we are developing innovative techniques to process the biomedical literature and extract information relevant to specific biological investigations. Biological experts routinely extract core information from the scientific literature using a manual process known as scientific curation. The aim of our activity is to improve the efficiency of this process by leveraging upon natural language processing technologies in a text mining system. There are two lines of investigation that we pursue: (1) finding information relevant for curation and present it in an adaptive interface, and (2) use sentence-similarity techniques to create interlinks across articles in order to allow a process of knowledge discovery.