MTGO-SC

MTGO-SC identifies functional gene modules in single-cell RNA-seq (scRNA-seq) data by constructing cluster-specific gene interaction networks that integrate network topology with external annotations to reveal cell-state–specific pathways.

Key Features:

• Adaptation for Single-Cell Data: Tailored for scRNA-seq to detect functional gene modules within individual cell clusters rather than averaging signals across bulk samples.
• Integration of External Annotations: Leverages Gene Ontology and Reactome pathway annotations to provide biological context for genes within networks.
• Network Topological Structure Utilization: Uses topological properties of gene interaction networks together with external annotations to isolate functional modules.
• Gene-Gene Interaction Networks: Represents direct interactions between genes within identified modules for fine-grained network analysis.
• Intermodule Interaction Networks: Enables examination of interactions between distinct functional modules to assess higher-level pathway relationships.
• Integration with Seurat: Works with the Seurat scRNA-seq analysis pipeline to operate on cluster definitions and processed single-cell data.
• Visualization and Filtering Functions: Provides functions to visualize gene networks and apply filters to network or module results.

Scientific Applications:

• Biomarker and Drug Target Discovery: Identifies candidate biomarkers and potential drug targets by locating functional modules associated with specific cell clusters.
• Cellular Heterogeneity Analysis: Defines distinct gene networks per cell cluster to investigate cellular heterogeneity and subpopulation-specific processes.
• State-Specific Pathway Analysis: Reveals cell-state–specific pathways and processes relevant to complex biological systems and disease states.

Methodology:

Constructs gene interaction networks for each cell cluster and identifies functional modules by integrating network topological information with external annotations (Gene Ontology, Reactome) while allowing configurable network extraction rules.