iBBiG

iBBiG applies an iterative bi-clustering algorithm to binary gene set profiles derived from gene set enrichment analyses to identify groups of gene sets coordinately associated with phenotypes across meta-analyzed gene expression datasets.

Key Features:

• Binary Gene Set Profiles: Transforms continuous gene expression profiles into binary gene set profiles by discretizing outputs from gene set enrichment analyses.
• Iterative Bi-clustering Algorithm: Uses an iterative bi-clustering algorithm to identify groups of gene sets coordinately associated with specific phenotypes across multiple studies without requiring predefined cluster counts.
• Robustness and Flexibility: Maintains performance in the presence of noise and discovers overlapping clusters of diverse sizes to capture complex gene set–phenotype associations.
• Application to Unmatched Samples: Optimized for meta-analysis of large numbers of datasets that may contain unmatched samples and data from multiple platforms.

Scientific Applications:

• Oncology — Breast cancer meta-analysis: Applied to integrate multiple breast cancer gene expression studies to extract associations between gene sets and phenotypes, including prediction of tumor metastasis within specific tumor subtypes.

Methodology:

Discretize gene set enrichment analysis outputs into binary gene set profiles, then apply an iterative bi-clustering algorithm to identify groups of gene sets coordinately associated with phenotypes across studies.