Evaluating Gene-Expression Clustering with Mutual Information
Author Information
Author(s): Priness Ido, Maimon Oded, Ben-Gal Irad
Primary Institution: Tel Aviv University
Hypothesis
Does the Mutual Information (MI) measure provide better differentiation among clustering solutions compared to traditional distance measures like Euclidean distance and Pearson correlation?
Conclusion
The study found that the MI measure significantly outperforms traditional measures in differentiating clustering solutions of varying quality.
Supporting Evidence
- The MI measure yields a more significant differentiation among erroneous clustering solutions.
- Clustering algorithms ranked oppositely when using different distance measures.
- The MI-based scores better differentiate among clustering solutions of different quality.
Takeaway
This study shows that using a special math tool called Mutual Information helps scientists group genes better than older methods.
Methodology
The study compared clustering solutions using MI, Euclidean distance, and Pearson correlation on several public gene expression datasets.
Limitations
The study primarily focused on specific datasets and may not generalize to all gene expression data.
Statistical Information
P-Value
0.0003
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website