HeteroGGM

HeteroGGM performs Gaussian graphical model-based heterogeneity analysis of high-throughput molecular and histopathological imaging data in an R package implementation.

Key Features:

• R package implementation: Provided as an R package enabling integration with R-based statistical workflows.
• Advanced Penalization Techniques: Employs penalization methods to manage high-dimensional data and to select significant conditional dependencies.
• Network-Based Analysis: Uses Gaussian graphical models to represent interconnections among variables and their individual properties as networks.
• Informative Summaries and Graphical Presentations: Produces summaries and graphical representations of network structures and inferred relationships.

Scientific Applications:

• Disease heterogeneity and biomarker discovery: Applied in biomedical research to uncover patient subtypes, identify novel biomarkers, and interpret molecular mechanisms driving variability using high-throughput molecular and histopathological imaging data.

Methodology:

Constructs Gaussian graphical models to represent conditional dependencies between variables (nodes represent variables such as gene expressions and edges denote conditional dependencies) and applies penalization techniques to control complexity and highlight significant connections.