Using Co-Authorship to Disambiguate Gene Names
Author Information
Author(s): Farkas Richárd
Primary Institution: Hungarian Academy of Science, Research Group on Artificial Intelligence
Hypothesis
A biologist refers to each particular gene by a fixed gene alias and this holds for the co-authors as well.
Conclusion
Co-authorship information can significantly improve the accuracy of gene name disambiguation in biomedical texts.
Supporting Evidence
- The study achieved a precision rate of 99.5% in gene disambiguation using co-authorship information.
- Disambiguation decisions were made in 85% of cases based on the inverse co-author graph.
- The methods introduced can be useful for other biomedical natural language processing tasks.
Takeaway
Scientists often use the same name for a gene in their papers, and by looking at who wrote those papers together, we can figure out which gene they mean.
Methodology
The study used graph-based semi-supervised methods to analyze co-authorship information for gene symbol disambiguation.
Potential Biases
Ambiguities in author names and the potential for misclassification of gene names due to spelling variants.
Limitations
The approach relies on a training set derived from manually disambiguated annotations, which may not cover all cases.
Participant Demographics
The study evaluated datasets from human, mouse, fly, and yeast organisms.
Statistical Information
P-Value
p<0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website