Predicting Protein Functions Using Literature and Interaction Data
Author Information
Author(s): Samira Jaeger, Sylvain Gaudan, Ulf Leser, Dietrich Rebholz-Schuhmann
Primary Institution: Humboldt-University Berlin
Hypothesis
Can conserved protein interaction graphs and literature mining improve the prediction of protein functions?
Conclusion
The method developed is highly reliable for predicting gene ontology annotations for poorly characterized proteins.
Supporting Evidence
- More than 80% of the GO annotations for proteins with highly conserved orthologs could be verified automatically.
- All predictions were correct according to the verifications from a trained curator.
- The method achieved 100% precision for the predicted GO annotations.
Takeaway
The researchers created a way to guess what proteins do by looking at how they interact with each other and what scientists have written about them.
Methodology
The study combined conserved protein interaction graphs with literature mining to predict protein functions.
Limitations
The method may not cover all proteins due to the limited availability of literature and the specificity of GO terms.
Statistical Information
P-Value
p≤0.01
Statistical Significance
p≤0.01
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website