Comparative Analysis of Protein Interaction Data Extraction Methods
Author Information
Author(s): Hena Jose, Vadivukarasi Thangavel, Devakumar Jyothi
Primary Institution: Jubilant Biosys Ltd.
Hypothesis
The study aims to compare the recall and precision rates of various natural language processing tools used for extracting protein interactions against manually curated data.
Conclusion
The evaluation revealed that the recall and precision rates of the NLP tools were lower than previously published scores when a normalized definition for interaction was applied.
Supporting Evidence
- The study compared NLP tool outputs with manually curated data to assess accuracy.
- Findings indicated that NLP tools often miss or incorrectly identify protein interactions.
Takeaway
This study looked at different computer programs that find protein interactions in research papers and found that they often make mistakes compared to human experts.
Methodology
The study compared data generated by selected NLP tools with manually curated protein interaction data to determine recall and precision rates.
Potential Biases
Potential bias may arise from the selection of NLP tools and the criteria used for comparison.
Limitations
The study did not conduct a comprehensive survey of all available NLP tools.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website