A large volume of protein-protein interactions has been identified, and information about such interactions is now readily available in online databases such as BIND. However, the information stored in current databases does not allow us to rank the biological validity of the interactions. It may be that interactions occurring under laboratory conditions do not actually occur in the living cell. A researcher trying to establish the quality of the interactions identified in a database could read the details of the experiments in each related scientific article, but this is labourious and time-consuming. If the number of relevant papers is high, it will be difficult or even impossible for a researcher to manually process all the articles to assess the value of the interactions.
Our goal is to develop automated biomedical information extraction systems that use discourse analysis methods to recognize and rank the quality of biomolecular interactions.
BioWebDoc is funded by a seed grant from the Waterloo Institute for Health Informatics Research (WIHIR).
Chrysanne DiMarco, Associate Professor, David R. Cheriton School of Computer Science
Alan Davidson, Associate Professor, Department of Biochemistry, University of Toronto
Xiaofen He (MMath, 2007)
Thesis topic: Using lexical chaining as a basis to rank the
biological validity of protein-protein interactions
Fred Kroon (PhD, ongoing)
Thesis topic: Mapping scientific communities using citation analysis
Xiaofen He
Gabe Musso, Department of Biochemistry, University of Toronto
Zhou Yu, Department of Biochemistry, University of Toronto
Chrysanne DiMarco
University of Waterloo, DC 1308
200 University Avenue West
Waterloo, Ontario, Canada N2L 3G1
519 888 4443
email: cdimarco@uwaterloo.ca