SGER: Facilitating Database Curation via Natural Language Processing

Columbia University, New York NY

Investigators

Abstract

This exploratory project investigates the effectiveness of natural language processing (NLP) techniques for facilitating curation and ontological development efforts of model organism databases. This research will enhance the infrastructure for research and education by increasing the productivity of curators and using computerized methods to assist in development of a standardized ontology of biomolecular relations. This project will address the following issues through a series of studies: a) what types of information are needed to actually reduce the work of curators, b) how should that information be presented, c) how much effort can actually be saved, and d) what level of performance is required? The studies will be performed utilizing curators and an existing NLP relation extraction system called GENIES that was shown to perform effectively. This project will also investigate how output from an NLP relation extraction system (GENIES) can be used to develop a standardized ontology of relations. This will involve exploration of how to organize and visualize the extracted relations for ontological work. The utility of a flexible XML-based visualization tool that allows users to view relations according to different dimensions, such as by agent, by target, or by frequency will be explored.