QUBIC2

QUBIC2 identifies condition-specific functional gene modules (biclusters) from microarray, bulk RNA-Seq, and single-cell RNA-Seq (scRNA-Seq) expression data to reveal condition-specific gene interactions.

Key Features:

• Left-Truncated Mixture of Gaussian Model: Models multimodality in zero-enriched expression data to accommodate abundant zero and low expression values typical of scRNA-Seq.
• Dropouts-Saving Expansion Strategy: Expands and optimizes functional gene modules by leveraging information divergency to manage dropout events in scRNA-Seq without compromising module integrity.
• Rigorous Statistical Testing: Applies a comprehensive statistical test to determine the significance of identified biclusters across organisms, including those lacking substantial functional annotations.

Scientific Applications:

• Condition-specific gene module discovery: Identifies biclusters that represent condition- or context-specific gene interactions and functional modules.
• Data-type applicability: Applicable to microarray, bulk RNA-Seq, and single-cell RNA-Seq (scRNA-Seq) expression datasets.
• Cross-organism analysis and benchmarking: Enables analysis across organisms including Escherichia coli and human datasets, and evaluation using simulated expression data.

Methodology:

QUBIC2 applies a left-truncated mixture of Gaussian model to assess multimodality in zero-enriched data, uses a dropouts-saving expansion strategy leveraging information divergency to optimize gene modules, and employs a comprehensive statistical test to assess bicluster significance.