GEMS

GEMS performs biclustering of microarray gene expression data to identify gene modules that exhibit coherent expression patterns across subsets of experimental conditions.

Key Features:

• Biclustering (co-clustering / two-way clustering): Identifies gene clusters that display consistent expression patterns across subsets of conditions to detect functionally related or co-regulated genes.
• Gibbs sampling: Applies a Gibbs sampling paradigm for bicluster mining to statistically explore the space of gene–condition biclusters.

Scientific Applications:

• Functional module discovery: Detects groups of genes participating in common biological pathways by finding coherent expression across condition subsets.
• Drug and disease response identification: Identifies genes whose expression responds to specific drugs or pathological states by isolating condition-specific biclusters.
• Regulatory module inference: Reveals co-regulated gene modules potentially controlled by limited sets of transcription factors through shared expression patterns.

Methodology:

Bicluster mining is performed by Gibbs sampling on microarray gene expression matrices after specification of biclustering criteria.