Predicting Protein Functions Using Hierarchical Data
Author Information
Author(s): Jiang Xiaoyu, Nariai Naoki, Steffen Martin, Kasif Simon, Kolaczyk Eric D
Primary Institution: Boston University
Hypothesis
Can integrating relational and hierarchical network information improve protein function prediction?
Conclusion
The proposed method significantly improves protein function prediction accuracy by utilizing the hierarchical structure of the Gene Ontology.
Supporting Evidence
- The method showed substantial improvements over standard methods like Nearest-Neighbor.
- Cross-validation indicated increased positive predictive value.
- In silico validation confirmed advantages suggested by cross-validation.
Takeaway
This study shows a new way to guess what proteins do by looking at how they are connected to each other and using a special tree structure that organizes information about their functions.
Methodology
A probabilistic framework was developed to integrate protein-protein interaction data with the Gene Ontology hierarchy for predicting protein functions.
Potential Biases
Potential biases may arise from the reliance on existing protein interaction data and the hierarchical structure of the Gene Ontology.
Limitations
The study focused on a specific organism (yeast) and may not generalize to other species.
Participant Demographics
The study involved proteins from the yeast Saccharomyces cerevisiae.
Statistical Information
P-Value
p<0.00001
Statistical Significance
p<0.00001
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website