OntoGene in BioCreative II
Author Information
Author(s): Fabio Rinaldi, Thomas Kappeler, Kaarel Kaljurand, Gerold Schneider, Manfred Klenner, Simon Clematide, Michael Hess, Jean-Marc von Allmen, Pierre Parisot, Martin Romacker, Therese Vachon
Primary Institution: University of Zurich
Hypothesis
How can novel protein interactions and experimental methods be effectively detected in scientific literature?
Conclusion
The study demonstrates that while no tool is fully reliable for automated annotations, some approaches perform competitively, making them useful for preliminary document inspection or as modules in curation environments.
Supporting Evidence
- The study describes approaches taken in the BioCreative competition to detect novel protein interactions and experimental methods.
- Results indicate that some proposed methods perform at a competitive level for preliminary document inspection.
- The study highlights the challenges of fully automated annotations in the context of biomedical literature.
Takeaway
This study shows how computers can help scientists find important information about proteins in research papers, even though they can't do it perfectly yet.
Methodology
The study used a combination of high-recall protein annotation, strict disambiguation steps, and pattern matching to identify protein interactions and experimental methods.
Limitations
The tools are not fully reliable for automated annotations and may miss some relevant interactions mentioned in full articles.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website