Biclustering of Gene Expression Data by Correlation-Based Scatter Search
Author Information
Author(s): Nepomuceno Juan A, Troncoso Alicia, Aguilar-Ruiz Jesús S
Primary Institution: University of Seville
Hypothesis
The proposed Scatter Search algorithm aims to find biclusters from gene expression data by using a fitness function based on linear correlation among genes.
Conclusion
The proposed algorithm successfully identifies biclusters with shifting and scaling patterns that are not detectable using traditional methods.
Supporting Evidence
- The algorithm was tested on three datasets: Yeast Cell Cycle, human B-cells lymphoma, and Yeast Stress.
- Biclusters with high average correlation were identified, indicating effective grouping of co-expressed genes.
- The proposed method outperformed traditional methods in detecting biclusters with shifting and scaling patterns.
Takeaway
This study created a new way to group genes that work together, helping scientists understand how genes behave under different conditions.
Methodology
The study used a Scatter Search algorithm to analyze three real datasets, focusing on finding biclusters based on gene correlations.
Limitations
The study may not account for overlapping among genes and the fitness function could be improved.
Statistical Information
P-Value
0.0069
Statistical Significance
p<0.001
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website