Related Experiment Videos
Analysis of expression profile using fuzzy adaptive resonance theory
Shuta Tomida1, Taizo Hanai, Hiroyuki Honda
1Department of Biotechnology, School of Engineering Nagoya University, Furo-cho, Chikusa-ku, Japan.
Bioinformatics (Oxford, England)
|August 15, 2002
Summary
This study introduces Fuzzy ART for gene expression clustering, outperforming other methods in analyzing Saccharomyces cerevisiae sporulation data. Fuzzy ART provides a robust and biologically validated approach for understanding gene functions.
Area of Science:
- Computational Biology
- Bioinformatics
Background:
- Gene expression clustering is crucial for understanding uncharacterized gene functions.
- Adaptive Resonance Theory (ART) offers a promising approach for effective clustering.
Purpose of the Study:
- To investigate and apply a clustering method based on Adaptive Resonance Theory (ART).
- To analyze time-series gene expression data during Saccharomyces cerevisiae sporulation.
Main Methods:
- Utilized Fuzzy ART, a variant of ART, for clustering gene expression profiles.
- Compared Fuzzy ART with hierarchical clustering, k-means, and self-organizing maps (SOMs).
- Validated clustering robustness with noisy data and defined a correctness ratio based on biological characterization.
Main Results:
- Fuzzy ART demonstrated superior clustering performance compared to hierarchical clustering, k-means, and SOMs.
- Mathematical validations confirmed Fuzzy ART's superior clustering reasonableness.
- Robustness was verified using noisy data, and biological validation confirmed its effectiveness.
Conclusions:
- Fuzzy ART is a highly effective method for gene expression data analysis.
- The approach provides biologically meaningful insights into gene functions.
- Software is available for broader application.