Gene Prioritization and Clustering Using Multi-View Text Mining
Author Information
Author(s): Yu Shi, Leon-Charles Tranchevent, Bart De Moor, Yves Moreau
Primary Institution: Katholieke Universiteit Leuven
Hypothesis
Incorporating multiple text mining models can improve the accuracy of disease-gene identification.
Conclusion
The multi-view text mining approach significantly enhances the performance of gene prioritization and clustering compared to single-view methods.
Supporting Evidence
- The multi-view approach demonstrated significantly better performance than single-view methods in both gene prioritization and clustering tasks.
- Integration of multiple controlled vocabularies improved the robustness of the results.
- Dimensionality reduction techniques like LSI were effective in enhancing prioritization performance.
Takeaway
This study shows that using different ways to look at gene information helps scientists find the right genes related to diseases better.
Methodology
The study used a multi-view text mining approach with various controlled vocabularies to index MEDLINE publications and evaluate gene prioritization and clustering.
Potential Biases
The reliance on specific controlled vocabularies may introduce bias in gene prioritization.
Limitations
The study does not address overlapping genes and the optimal number of clusters in clustering tasks.
Statistical Information
P-Value
2.67e-004
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website