Mining physical protein-protein interactions from the literature
2008

Mining Protein-Protein Interactions from Literature

Sample size: 740 publication Evidence: moderate

Author Information

Author(s): Huang Minlie, Ding Shilin, Wang Hongning, Zhu Xiaoyan

Primary Institution: Tsinghua University

Hypothesis

There is a need for effective text-mining tools to extract physical protein-protein interactions from the literature.

Conclusion

The study presents a text-mining framework that effectively extracts physical protein-protein interactions, ranking among the top performers in the BioCreative 2006 evaluation.

Supporting Evidence

  • The method achieved a precision of 75.07% and a recall of 81.07% in filtering irrelevant articles.
  • In identifying protein mentions, the method had a precision of 34.83% and a recall of 24.10%.
  • The profile-based method was competitive, achieving a precision of 36.95% on the SwissProt-only subset.

Takeaway

This study created a tool that helps scientists find out how proteins interact with each other by reading lots of scientific papers quickly.

Methodology

The study used a text-mining framework that included article filtering, protein mention identification, normalization to molecule identifiers, and extraction of protein-protein interactions.

Potential Biases

The reliance on training data that may not represent the full range of literature can introduce bias.

Limitations

The method struggles with protein name normalization and requires evidence to be present in the same sentence for interaction extraction.

Statistical Information

P-Value

0.02

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1186/gb-2008-9-s2-s12

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication