GaneSh

GaneSh performs model-based coclustering of genes and experimental conditions using a Bayesian framework to identify biologically meaningful coexpression modules in large-scale gene expression datasets.

Key Features:

• Model-Based Clustering Approach: Uses model-based clustering grounded in statistical principles for analysis of complex gene expression data.
• Simultaneous Gene and Condition Clustering: Coclusters genes and experimental conditions within a single analysis framework to capture gene–condition interactions.
• Bayesian Framework with Gibbs Sampling: Applies Bayesian inference with a Gibbs sampling procedure to iteratively update cluster assignments and to explore the posterior distribution, including cases where the posterior is strongly peaked around a limited number of equiprobable clusterings.
• Graph Spectral Method for Fuzzy Clustering: Extracts fuzzy, overlapping clusters from a weighted graph representation of genes using a graph spectral method, identifying core tightly coexpressed gene sets and overlapping partial coexpression relationships.
• Biological Significance and GO Annotation Analysis: Validates identified clusters through Gene Ontology (GO) annotation analysis demonstrating biological relevance of local maxima.
• Scalability for Large-Scale Datasets: Tailored for large-scale gene expression datasets such as microarray experiments.

Scientific Applications:

• Large-Scale Gene Expression Analysis (microarrays): Identifies coexpression modules and condition-specific patterns in microarray and other large-scale expression datasets.
• Systems Biology: Detects coregulatory relationships to inform network and systems-level analyses.
• Functional Genomics: Supports functional genomics studies by discovering biologically coherent gene clusters validated by GO annotation.
• Regulatory Mechanism and Interaction Discovery: Aids in uncovering complex regulatory mechanisms and potential genetic interactions by revealing core and partial coexpression networks.

Methodology:

Bayesian model-based coclustering using Gibbs sampling to iteratively update cluster assignments; graph spectral methods on a weighted gene coexpression graph to extract fuzzy overlapping clusters; Gene Ontology (GO) annotation analysis for biological validation.