TopologyGSA

TopologyGSA performs graphical-model-based pathway analysis of gene expression data to detect pathway-level and component-level changes by comparing mean and concentration matrices while accounting for heteroscedasticity.

Key Features:

• Graphical model-based analysis: Employs graphical models to incorporate the dependence structure among genes within pathways.
• Mean-matrix comparison test: Implements a statistical test comparing mean matrices across experimental groups to detect overall changes in pathway gene expression.
• Concentration-matrix comparison test: Implements a statistical test comparing concentration matrices to assess changes in the strength of interactions between genes within pathways.
• Component-specific analysis: Enables examination and comparison of individual pathway components (cliques) to identify deregulated substructures.
• Heteroscedasticity consideration: Accounts for heteroscedasticity in differential expression tests.

Scientific Applications:

• Pathway-level differential analysis: Performs pathway-level surveillance using a priori pathway definitions such as those from the Kyoto Encyclopedia of Genes and Genomes (KEGG).
• Component-level deregulation detection: Identifies specific pathway components involved in deregulation to support studies of disease mechanisms or treatment effects.

Methodology:

Uses graphical models, statistical tests comparing mean matrices and concentration matrices, component (clique) analysis, and heteroscedasticity-aware differential testing.