Improving Gene Interaction Analysis with Seeded Bayesian Networks
Author Information
Author(s): Djebbari Amira, Quackenbush John
Primary Institution: Dana-Farber Cancer Institute and Harvard School of Public Health
Hypothesis
Can prior network structures improve the ability of Bayesian Network analysis to learn gene interaction networks from microarray data?
Conclusion
Using network seeds significantly enhances the ability of Bayesian Network analysis to identify gene interactions from gene expression data.
Supporting Evidence
- The use of prior network seeds improved recovery of known interactions.
- Bootstrapping provided confidence estimates for network features.
- Networks constructed with high confidence edges had a low false-positive rate.
Takeaway
This study shows that using information from previous research helps scientists better understand how genes interact with each other.
Methodology
The study used Bayesian Network analysis with prior network structures derived from literature and protein-protein interaction data, applying bootstrapping for confidence estimation.
Potential Biases
Using prior knowledge may bias the results towards known interactions.
Limitations
The analysis focused on specific pathways and may not generalize to all gene interactions.
Participant Demographics
The study analyzed pediatric samples of Acute Lymphoblastic Leukemia and Acute Myeloid Leukemia.
Statistical Information
P-Value
1.18 × 10-6
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website